What's New
Customers can now call your business on WhatsApp and have a voice agent answer. You use the WhatsApp Business number you already have, so there is no extra phone number to buy.

More Detail

Before you start

• The voice agent must be fully set up. Until it is, the page shows "Finish configuring this voice agent before connecting a WhatsApp number." and the number switches stay off.

• The number must run on WhatsApp's Cloud API. A number that is still in the WhatsApp Business app can't take calls (see Note).

• Your wallet needs a balance above zero. If it is empty, switching a number on shows: "Your wallet balance is empty, so calls on this number could not be enabled. Please top up your wallet and try again."

• Sub-account users only see this option if they have permission to manage integrations.

Step 1: Open the WhatsApp page

1. In the sidebar, click Voice Agents and open your agent.

2. Click the Integrations tab, next to Inbound, Outbound and Call Logs. If the voice agent hasn't been created yet, the tab is disabled and shows "Please create a voice agent first".

3. On the WhatsApp card ("Let customers call this voice agent on your WhatsApp Business number."), click Connect.

Step 2: Authorise with Facebook

1. The page shows Not connected and a Three steps to go live guide: Authorise Meta, Pick your numbers, Enable and go live. To see what access is requested, click See the five permissions Meta will request.

2. Click Authorise with Facebook. Sign in with the account that manages your WhatsApp Business numbers and choose the numbers to share.

3. You come back to the same page and see one of these messages:

   • "WhatsApp number authorised successfully." when it worked.

   • "Access Denied" if you cancelled on Facebook.

   • "WhatsApp number authorisation failed." if something went wrong.

Step 3: Turn a number on

1. The top card now reads Connected through Meta, with a count such as "1 number synced · 0 live". Its status shows Nothing routing until a number is on.

2. Under Numbers ("Enabled numbers ring this voice agent."), find your number. Use Search numbers if you have many.

3. Flip the number's switch on. It shows Updating…, then Live, with "WhatsApp calls are now answered by this voice agent."

4. Each number can be answered by only one voice agent. If another agent already uses it, you see "This number is already connected with Voice Agent [agent name]."

What callers experience

• Callers tap the call button in your WhatsApp chat and the voice agent answers. It uses the same instructions, knowledge and settings as on phone calls.

• WhatsApp calls follow the same rules as phone calls:

  • the agent must be active

  • calls are only taken during its available hours

  • blocked numbers are refused

  • your concurrent-call limit applies

  Refused calls end without being answered and are not charged.

• If someone texts the number and it isn't also connected to a chat agent, they get one automatic reply (at most once an hour per person): "Thanks for reaching out! This number is answered by a voice assistant, so it can't reply to messages. Please tap the call button to talk to us."

Managing and turning off

• To stop calls, flip the number's switch off. It shows Paused and "WhatsApp calls are no longer answered by this voice agent."

• To remove a number:

  1. Click the ⋮ menu on its row and choose Delete number.

  2. Confirm in the Delete number? dialog. It reads "[number] will stop taking calls for this voice agent and will be removed from this integration. You can add it again by authorising through Meta."

• The same ⋮ menu also has Copy number.

• To add more numbers, click Add another number.

• To renew access, choose Re-authorise with Meta from the top card's ⋮ menu.

• Deleting the voice agent also switches off calling on its WhatsApp numbers.

Impact
Affects Agency Users and Sub-accounts using Voice Agents. Customers can reach your voice agent from inside WhatsApp, with no new number to buy.

Note

• WhatsApp calls use your standard voice per-minute rate from the wallet. There is no extra carrier charge. Sub-account calls are paid from the agency wallet, the same as other inbound calls.

• If the wallet runs out, the number is switched off after the current call ends and shows Paused. Top up, then switch it back on.

• Numbers still running in the WhatsApp Business app can't take calls. Switching one on shows: "This number is still in use in the WhatsApp Business app. WhatsApp calling is only available for numbers running on the Cloud API."

• If calling isn't available on your account, the page shows "WhatsApp calling for voice agents is not available yet. Please contact support."

(Chat) WhatsApp Coexistence: Keep Using the WhatsApp Business App Alongside Your Chat Agent

What's New
You can now connect a number that runs in the WhatsApp Business app without removing it from the app. Your chat agent replies to customers, you can still reply from your phone, and the agent steps back while you're talking.

More Detail

Step 1: Choose how to connect

1. In the sidebar, click Chat Agents and open your agent.

2. Click the Integrations tab, then the WhatsApp card.

3. You now see two buttons:

   • Migrate my number to the API. Hover text: "Moves the number to the Cloud API. It must be removed from the WhatsApp Business app first." This is the standard connection and hasn't changed.

   • Keep using the Business app. Hover text: "Keeps the number working in the WhatsApp Business app while the agent replies alongside you."

4. A note under the buttons explains that keeping your number in the app only works while the number is still in the app. Once a number has been moved to another service, it can't go back.

Step 2: Connect the number

1. Click Keep using the Business app. You are taken to Facebook's WhatsApp signup to link your WhatsApp Business app number.

2. When you finish, you return to the WhatsApp page and see "WhatsApp number authorised successfully."

3. Under Numbers, flip the number's switch on. It shows Live and "WhatsApp number connected successfully."

How to tell a Business app number apart

• These numbers show a Business app label on their row. Its hover text reads: "This number is still live in your WhatsApp Business app. To stop using it here, disconnect from the app under Settings > Account > Business Platform."

Day to day

• The chat agent answers incoming messages as usual. Its replies also appear in your WhatsApp Business app.

• When you reply from the app, your message is saved in the same conversation under Conversations → Inbox.

• The conversation's Chat Agent / Human switch then moves to Human, and the agent stops replying so the customer doesn't get two answers.

• When you stop replying for a short while, the switch moves back to Chat Agent and the agent starts answering again.

• Conversations handed to a person another way stay with that person and are not switched back automatically. For example, the customer asked for a human, or someone turned on Human mode in the Inbox.

Disconnecting

• To stop using a Business app number here, disconnect it inside the WhatsApp Business app under Settings > Account > Business Platform.

Impact
Affects Agency Users and Sub-accounts using WhatsApp with Chat Agents. Businesses already running on the WhatsApp Business app can add an AI agent without giving up the app or their number.

Note

• This only works for numbers still running in the WhatsApp Business app. If you've already connected a number the standard way, the button's hover text reminds you that "A number already on the Cloud API cannot move back."

• Chats from before you connected stay in the WhatsApp Business app. They are not copied into Stammer.

• Some countries are not yet supported for this connection method.

• WhatsApp calling for voice agents isn't available on Business app numbers.

• If this option isn't available on your account, choosing it shows "Keeping your number in the WhatsApp Business app is not available yet. Please contact support."

• Numbers you have already connected are not affected.

Conversations: Inbox, Analytics and Summary Settings in One Place

What's Changed
Analytics and Summary Settings now live inside each chat agent's Conversations tab, next to the Inbox. Conversations, reporting and summary setup are all in one place.

More Detail

• Inbox is still the default view when you open Conversations.

• Each sub-tab has its own link, so you can bookmark or share it. Old Analytics and Summary Settings links still work.

• The Leads view now matches the rest of the dashboard. Long links and file names are shortened, and you can still open or copy the full value.

• Text in the Model dropdown now matches the size of the other fields.

Impact
Affects all users managing Chat Agents. You can review, analyse and set up conversations without leaving the page.

Note
Analytics and Summary Settings work the same as before. Only their location has changed.

(Chat) Redesigned Chat Agent Cards

What's Changed
The Chat Agents list has new cards. Each card shows a small preview of the agent's chat widget in its own colours and icon, so agents are easier to recognise.

More Detail

• Each card shows the agent's Chats and Messages counts.

• Labels such as Disabled, Pinned to Top, Agency Default, SubAccount Default, Internal Support and the MasterChat icon are easier to spot.

• The folder tabs and the card's ⋮ menu have been refreshed to match the new design.

Impact
Affects all users with Chat Agents. You can find and tell agents apart more quickly, especially in accounts with many agents.

Note
Cards no longer show the trained character count or the model name. You can still find both inside each agent.

Billing → Billing Usage & Settings → Adding Cards That Require Bank Verification

What's Changed
Some banks ask for an extra security check (3D Secure) when you add a card. This is common for European cards. Before, these cards failed with "Payment Method has not been created successfully." Now your bank's check opens and the card saves once you approve it.

Impact
Affects all users adding a payment method, especially customers whose bank requires card verification. They can now add a card without contacting support.

Note
Cards that don't need verification are added the same way as before.

Improved Translations

What's Changed
We fixed many translations across all 11 non-English languages. Some screens showed English instead of your chosen language, and some showed a translation that belonged to a different label.

Impact
Affects users who use the dashboard in a language other than English. More screens now show the correct text in your language.

Note
We are still working on translations, and more fixes will come in future releases.

Live Support Option

We've added a new live support option to help you out!

📞 Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with the Stammer team. Available Monday to Friday from 11:30am to 12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

What's New We have launched a new plan structure built around three plans, Standard, Pro (White Label), and Enterprise, each with tiers you can scale up as you grow. Yearly billing saves 20% on every tier.

More Detail

Standard is for running AI chat and voice agents for your own business. It includes 1 AI chat agent and 1 AI voice agent, up to 3 users, live chat with human handoff, lead collection, scheduling, basic integrations (web chat, Facebook, Instagram, WhatsApp, and API), and email support. Pick the tier that matches your volume:

Go past your included usage and you simply pay as you go at $0.002 per reply and $0.10 per voice minute. Standard comes with a 7-day free trial. White labelling is not included.

Pro (White Label) is for agencies reselling AI agents to clients. It is fully white labeled on your own domain, logo, and colours, with unlimited sub-accounts, the ability to resell AI usage at your own prices, all AI models, all integrations, MCP support, bring-your-own LLM API keys, white label API access, advanced analytics, 1-on-1 onboarding, and daily office hours. Tiers scale by agent count:

Pro comes with a 14-day free trial.

Enterprise is custom-built AI automation at enterprise scale, with unlimited chat and voice agents and no usage caps. Pricing is tailored to you, so book a call at https://stammer.ai/enterprise.

Alongside the plans themselves:

• The pricing page has been redesigned so tiers, what is included, and custom pricing are easier to compare at a glance.

• Every feature on a plan card now carries a tooltip explaining exactly what it means, so there is no guessing what you are paying for.

• Switching plans or tiers mid-cycle has been reworked so credits, billing cycles, and free trial handling are calculated correctly on an upgrade or downgrade.

• Free trials now run to fixed terms and can be cancelled directly.

Impact Affects all Agency Users and Sub-accounts on paid plans, and anyone signing up from now on. Clearer pricing, a tier that fits your actual volume rather than a jump between two extremes, and plan changes that do the right thing with your remaining credits.

Note Older legacy plans remain available to accounts already on them and are no longer listed for new signups. On Standard, unused monthly credit does not roll over into the next cycle.

(Chat) Page-Aware Chat Widget

What's New Your chat agent can now tell which page a visitor is on and use that page's content in its answers, including a page-specific opening message instead of the same generic greeting everywhere.

More Detail

• Turn it on per chat agent in your agent settings. It is off by default, so nothing changes until you enable it.

• The widget matches the visitor's current page against the content you have already scraped and trained on, so there is no extra setup on your website.

• When a match is found, the agent can reference what is on that page in its replies, and the opening message becomes specific to that page (for example, a product page can open with a line about that product).

• Page-specific opening messages are generated once per page and cached, then refreshed automatically whenever that page's scraped content is updated.

Impact Affects all Agency Users and Sub-accounts using the chat widget. Visitors get answers and greetings that match what they are actually looking at, which lifts engagement on product and landing pages.

Note Disabled by default and opt-in per agent. Page content is only used where it matches content you have already trained the agent on.

(Chat) Re-Engagement Nudge When a Conversation Goes Quiet

What's New Chat agents can now send a follow-up message when a visitor stops replying mid-conversation, bringing distracted visitors back to the chat.

More Detail

• Enable nudges per chat agent in your agent settings. Off by default.

• Set how long the conversation must be quiet before a nudge is sent, anywhere from 1 to 60 minutes.

• Write your own nudge message, and choose how many nudges a single conversation can receive (up to 3) so it never becomes intrusive.

• Nudges are marked separately in the conversation so you can tell them apart from normal agent replies in your inbox.

Impact Affects all Agency Users and Sub-accounts using chat agents. Recovers conversations, and the leads in them, that would otherwise be abandoned.

Note Disabled by default and opt-in per agent, with a hard cap on how many nudges a conversation can receive.

(Voice) Scheduled Available Hours

What's New Voice agents can now be limited to a schedule, so the agent only answers calls during the hours you choose instead of 24/7.

More Detail

• Open your voice agent settings and find the availability section, which works the same way as the existing scheduled hours for human handoff.

• Choose "All time" to keep today's always-on behaviour, or "Scheduled" to set specific hours.

• Pick your timezone, then set each day of the week individually: on or off, all day, or a specific time range for partial availability.

• Calls that arrive outside the configured schedule are not handled by the agent.

Impact Affects Agency Users and Sub-accounts using Voice Agents. Ideal for businesses that want the agent purely as an after-hours solution, and prevents the agent from answering when your own team is already on the phones.

Note "All time" remains the default, so existing voice agents keep answering around the clock until you change the setting.

Exclude URL Patterns When Scraping a Website

What's New You can now tell an agent which parts of a website to skip when scraping, so pages like cart, checkout, login, and account never get crawled or trained on.

More Detail

• Add exclusion patterns in the website scraping settings, alongside the website URLs you already enter.

• Patterns use simple wildcards, for example */cart/*, so you do not need to list every page.

• Exclusions apply both to URLs you enter directly and to links the crawler discovers while working through the site.

• Excluded pages are never fetched at all, so they cost nothing in scraping time, storage, or training.

• The setting is shared by both chat agents and voice agents.

Impact Affects all Agency Users and Sub-accounts training agents from a website. Keeps noisy, private, or irrelevant pages out of your knowledge base and reduces wasted training storage.

Note Exclusions apply going forward. Pages already scraped before you add a pattern are not removed automatically.

More Customisable Email Templates

What's New Account Settings → Email Templates now offers several more templates you can word yourself, instead of just Lead Collection.

More Detail

• New templates include Human Handoff, Unanswered Question, Sub-account Welcome, Account Activation, Conversation Summary, and voice phone number recharge notifications.

• Each template shows its own name, description, default subject and body, and the exact variables available to it, so the editor always matches what the real email sends.

• Variables now work in the subject line as well as the body, so subjects can include dynamic details.

• Saved templates are validated against that template's known variables, so a typo is caught before it reaches your customers.

• Templates you have not customised continue to use our defaults.

Impact Affects Agency Users white-labelling the product. More of the emails your customers receive can now carry your own wording and tone.

Note Which templates you see depends on the features enabled on your account.

Redesigned Inbox

What's Changed The inbox has been rebuilt with a cleaner, more consistent interface across conversations, contacts, and leads.

More Detail

• Refreshed conversation list and conversation detail views with a consistent visual style.

• Contact names, avatars, and profile pictures now render correctly across channels.

• Timestamps are shown on agent responses as well as incoming messages.

• Leads are now available directly in the global inbox.

Impact Affects all users who manage conversations in the inbox. Faster to scan, easier to follow a conversation, and consistent no matter which channel the message arrived on.

Redesigned Meta Integrations Tab

What's Changed The Meta integrations area has been redesigned into a clearer dashboard for connecting and managing WhatsApp, Instagram, and Facebook Messenger.

More Detail

• A single dashboard view of your connected Meta channels and their status.

• Clearer connection and configuration flows, with confusing steps and dead ends cleaned up.

• Fixed tab redirection so you land on the right place after connecting or updating a channel.

• WhatsApp configuration is now available for voice agents as well.

Impact Affects Agency Users and Sub-accounts using WhatsApp, Instagram, or Facebook Messenger. Less guesswork when connecting a channel and a clearer picture of what is already live.

Attachments No Longer Lost After Human Handoff on WhatsApp, Instagram, and Messenger

What's Changed Images and files a customer sends after a conversation is handed off to a human agent now reach the inbox instead of disappearing.

More Detail

• Previously the agent saw a completely blank message with no sign that anything had been attached, and the file could not be recovered afterwards.

• This now works the same way it already did while the AI was answering, across WhatsApp, Instagram, and Facebook Messenger.

Impact Affects Agency Users and Sub-accounts using Meta channels with human handoff. Photo evidence, screenshots, and invoices sent during an escalation now arrive intact, so your team no longer has to ask the customer to resend.

Undelivered Meta Messages Are Now Detected and Retried

What's Changed We now read delivery status updates from WhatsApp, Instagram, and Facebook Messenger, so a message that was accepted but never actually delivered is caught rather than counted as a success.

More Detail

• Outbound messages on all three Meta channels are now tracked through to delivery.

• Failed or undelivered messages are recorded against the conversation, including the reason.

• Transient failures are retried automatically, and permanent failures are surfaced instead of being silently dropped.

Impact Affects Agency Users and Sub-accounts using Meta channels. You can now tell the difference between a message that landed and one that quietly never arrived, which is often caused by the connected business account's own setup.

Reliable Tool and Function Calls Across All AI Models

What's Changed Custom functions and tools now behave consistently no matter which AI model your agent runs on.

More Detail

• If one tool in a turn fails, the agent now reports the failure back to itself and continues the reply instead of ending the turn with an error. Previously, work that had already completed, such as a booking created or a lead saved, could be left applied while the customer got nothing back.

• When a model requests several tools in the same round, all of them now run. Previously, on some models only the first ran and the rest were dropped with no error, so a reply could describe work that never happened.

• Booking and calendar failures caused by a missing required field are now handled and reported rather than failing hard.

Impact Affects all Agency Users and Sub-accounts using custom functions, tools, or integrations in their chat agents. Fewer dead-end conversations and far fewer half-completed actions.

Dashboard Pages Flashing Through Multiple Loading Skeletons

What's Changed Dashboard pages now show one continuous loading state instead of two or three different placeholders in sequence.

More Detail

• Previously the layout visibly rebuilt itself two or three times before real content appeared, which read as the page stuttering.

• The skeleton now appears once, holds its shape, and is replaced by real content with no jumping.

• The behaviour is consistent across dashboard pages rather than fixed page by page.

Impact Affects all users. Navigation feels noticeably smoother and less glitchy, especially on slower connections.

Faster, Lighter Chat Widget on High-Traffic Sites

What's Changed The chat widget now reuses its configuration across page views instead of re-fetching it on every single page load.

More Detail

• On multi-page websites, a visitor browsing several pages now triggers at most one configuration request per cache window rather than one per page.

• Repeat requests after the cache expires are answered without re-sending the full configuration.

• A per-agent control lets us cap or switch off a single widget if it is causing problems, without affecting anyone else.

Impact Affects all Agency Users and Sub-accounts embedding the chat widget, especially on high-traffic websites. The widget loads faster for visitors and puts far less load on the platform.

Note Configuration changes you save can take a short time to reach live visitors while the cache refreshes.

Sub-Accounts → Edit Form and Dashboard Wallet Display

What's Changed Agencies can now save changes to a sub-account without being blocked by fields they cannot edit, and directly billed sub-accounts no longer show a misleading wallet balance.

More Detail

• The edit form now only validates fields the agency can actually change. Read-only fields keep their current values and never block a save.

• The Dashboard sub-accounts table now shows "Directly Billed" in the wallet columns for directly billed sub-accounts, matching what the Sub Accounts page already showed.

Impact Affects Agency Users managing sub-accounts. Edits no longer fail on a greyed-out field, and the dashboard no longer implies a directly billed sub-account has a zero wallet.

Graceful Handling of Third-Party Failures

What's Changed Temporary outages and expired credentials from connected services are now caught and reported to the account that owns them, rather than surfacing as a hard error.

More Detail

• Temporary failures are retried automatically.

• Permanent problems, such as an expired API key or a social channel token invalidated after a password change, are recorded against the affected account and surfaced so you can fix them.

• Lead extraction now retries on a temporary provider blip instead of failing outright.

Impact Affects all Agency Users and Sub-accounts using their own API keys or connected social channels. You get a clear signal about what needs reconnecting instead of unexplained failures.

UI Copy and Display Fixes

What's Changed A group of small copy and formatting corrections across the dashboard.

More Detail

• The Tools tab and its page header now use one consistent name instead of two.

• Scraped-URL dates now display correctly in Spanish.

• The knowledge base re-train button now names the correct agent type.

• Fixed a truncated phone number display.

• Custom CSS now applies correctly to the chat widget and chat bubble.

Impact Affects all users. Small things, but they remove confusion and make the product read correctly in every language we support.

Live Support Option

We've added a new live support option to help you out!

(NEW) 📞 Daily Support Calls Need help getting started or figuring something out? Join our new daily 30-minute calls with the Stammer team. Available Monday to Friday from 11:30am–12:00pm PST. Link to join: https://stammer.ai/daily-support-call

What's New

Signing up no longer drops you into an empty dashboard with a pricing page attached. New agencies and their sub-accounts are now walked through a guided setup that builds, trains, and hands them a working Chat Agent and Voice Agent, all before a plan is ever selected. Paste a website URL and we assemble the whole thing for you: brand, knowledge, prompt, both agents. Then you talk to your chat agent, call your voice agent, and decide what to pay for once you have already seen it work on your own business.

More Detail

• Built from your website. Enter a URL. We pull your logo, brand colours, business summary, training sources, and a starter prompt. Or choose "Continue Manually" and enter it yourself.

• Everything stays editable. Your knowledge lands in Files, Text, Website URL, and Q&A sections. Add, edit, remove, or skip anything.

• Two agents, one setup. "Continue & Train" creates and trains a Chat Agent and a Voice Agent together, from the same prompt and knowledge.

• Then you try them. Chat with your Chat Agent. Place a real test call to your Voice Agent. Both are trained on your own business.

• $5 credit to try it on us. Agencies claim a signup credit before the demo, so the first conversations and test call are covered. Skipping the step still applies the credit.

• The plan picks itself. Tell us if you are a Business Owner, Agency, or Enterprise, and the packages page opens with the right plan preselected.

• Sub-accounts get the same journey, in your brand. Your domain, your branding, your packages, Stammer branding hidden. Free agent quotas come from your free-trial settings. The credit and account-type steps are skipped.

Impact
Affects new Agency Users and new Sub-accounts.

New users reach the pricing page with a trained chat agent and voice agent already working, so upgrading is a confirmation rather than a leap of faith. Agencies get the same flow selling their own packages, under their own brand, without running a demo call for every signup.

Note
Off by default. Agency and sub-account onboarding are separate admin settings, so you can enable one first and add the other later. Until you switch them on, your signup experience is unchanged.

If an agency allocates zero chat and zero voice agents in its free-trial settings, its sub-accounts skip the build-and-try flow and go straight to plan selection.

(Chat) Voice Notes on Instagram, WhatsApp and Messenger

What's New

Customers can now send a voice note on Instagram, WhatsApp, or Messenger and your agent will answer what they actually asked. Previously voice notes were answered blind on Instagram and Messenger, and ignored entirely on WhatsApp.

More Detail

• Works automatically on any Chat Agent connected to Instagram, WhatsApp, or Messenger. There is nothing to switch on.

• An inbound voice note is transcribed by our voice engine and the transcript becomes the customer's question, so the agent replies to the real request.

• Transcription is metered on the actual length of the audio and debited from the wallet that pays for that agent. For sub-accounts on Direct Agency Billing, the charge goes to the agency's wallet instead.

• Before transcribing, we check the wallet can cover the cost and attempt an auto-recharge if it is enabled. If the balance still cannot cover it, the voice note is skipped rather than running up an unpayable charge.

• If a voice note cannot be understood, the customer gets a clear reply asking them to resend it or type their message, instead of silence or a guessed answer.

• Voice note charges appear in your wallet transaction history under their own "Voice Note Transcription" reason, labelled with the channel it came from.

Impact

Affects Agency Users and Sub-accounts running Chat Agents on Instagram, WhatsApp, or Messenger. Voice-first customers finally get real answers, and the cost is visible line by line in the wallet.

Note

Each voice note is charged exactly once, even if the same message is delivered again.

Conversations > Media Attachments: Filenames, Previews and Spacing

What's New

File attachments in the conversation view now show the customer's original filename and can be previewed inside the dashboard instead of downloading them first.

More Detail

• Open any conversation in the dashboard inbox and click a file attachment to preview it in place. PDFs, images, and other previewable types open directly in the dashboard.

• Download is still available on every attachment, and file types that cannot be previewed keep their download card as before.

• File cards now show the exact filename the customer sent, including the extension, truncated only where the layout needs it. Previously they showed an internal generated name.

• Attachments and message bubbles now have consistent vertical spacing, making long threads easier to scan.

Impact

Affects all users reviewing conversations in the dashboard. Reviewing a customer's document no longer means downloading it and guessing which file it was.

Note

Attachments received before this release show the closest match we can recover from the stored filename. Image rendering is unchanged.

(Chat) Longer Conversations: Prompt Size Limit Removed

What's Changed

The cap on how large a chat prompt could grow has been removed. Long conversations and large knowledge context no longer fail partway through.

Impact

Affects all users running Chat Agents on our managed AI. Conversations now run up to the AI model's real limit rather than stopping early, agents can draw on more history and knowledge in each reply, and agency owners no longer receive "token limit exceeded" emails for conversations that should have worked.

Note

Usage is billed per token as before, so longer conversations consume proportionally more from your wallet. Agents using your own AI key were never affected by this cap.

Agency Settings > Custom Domain: Accurate Deployment Status

What's Changed

The custom domain status now tells you what actually happened. A deployment that is still running, or one that failed, is no longer reported as successful.

More Detail

• A running deployment now shows its own "in progress" state rather than a success message.

• A failed deployment shows the real reason it failed, with a prompt to try again.

• A deployment that stalls far beyond a normal run is now reported as failed instead of appearing to run forever.

• The success message only appears when the domain genuinely deployed.

Impact

Affects Agency Users setting up a custom domain. You now find out immediately when a domain did not deploy, and why, instead of being told it worked while the domain is down.

Billing > Auto-Recharge Emails No Longer Sent to Directly Billed Clients

What's Changed

Sub-accounts on Direct Agency Billing no longer receive Chat Agent or Voice Agent billing and auto-recharge emails, including the "Auto Recharge for Chat Agents Disabled" notice.

More Detail

• Directly billed clients have no payment method or billing settings of their own, so these emails were confusing and pointed to a page they cannot open.

• Chat Agent and Voice Agent billing emails now follow the same rule wallet and storage emails already followed.

• When an agency transfers a chatbot to a directly billed client, the capacity check is now attributed to the agency that is actually paying, so any resulting notice reaches the agency admin instead of the client.

Impact

Affects Agency Users using Direct Agency Billing and their Sub-accounts. Clients stop receiving alarming billing emails about operations that succeeded, and agencies get the notices that are genuinely theirs to act on.

Sub-accounts: Correct Resource Allocation from Free Trial to Paid Plan

What's Changed

Corrected how agency resource counters track sub-accounts as they move from free trial to a paid plan and when accounts are disabled or reset.

More Detail

• Concurrent call slots and voice knowledge base slots are now returned to the agency pool when a sub-account is disabled or reset, instead of staying permanently consumed.

• Sub-account permissions are now updated automatically when they subscribe to a paid agency package, so they get their plan's entitlements without an agency admin editing them by hand.

• Concurrent call allocation is now counted correctly from the moment a free trial starts, so the agency's available capacity reflects reality.

Impact

Affects Agency Users with Sub-accounts. Agency capacity numbers now match what is actually in use, preventing over-allocation beyond your real limits and removing manual permission fixes after a sub-account upgrades.

Conversations > Export: CSV Now Respects the Selected Date Range

What's Changed

Exporting conversations with a date filter now includes only the messages sent inside the selected range.

More Detail

• Previously, any conversation that was active during your date range had its entire message history written to the CSV, including messages from months earlier.

• Which conversations appear in the export is unchanged: a conversation that received a message in your range is still included, even if it started earlier.

Impact

Affects all users exporting conversations. Filtered exports now contain only the messages you asked for, so reports and audits match the period you selected.

(Chat) Fixed Replies Cutting Off When an Agent Used a Tool

What's Changed

Fixed a fault where a streaming reply could start and then break mid-response whenever the agent needed to look something up or run one of its functions.

Impact

Affects all users running Chat Agents that use knowledge lookups or custom functions. Replies that use a tool now complete reliably instead of stopping partway.

(Chat) Agents Can Handle Multiple Actions in a Single Turn

What's Changed

Agents can now run several actions in one turn when a question needs more than one lookup, rather than being limited to one at a time. Usage across every step of the turn is counted correctly.

Impact

Affects Agency Users and Sub-accounts running Chat Agents with custom functions or integrations. Multi-part questions get answered in a single turn, and the wallet charge reflects the full turn rather than undercounting it.

Note

This is rolling out gradually and is off by default per agent. Existing agents keep their current behaviour until it is enabled for them.

Faster Dashboard and Chat Widget Loading

What's Changed

Speed improvements across the dashboard, the public chat widget, and conversation responses.

More Detail

• Chat widget configuration and lead-capture fields are now served from cache instead of being looked up on every single visitor load.

• Conversation attachments are served by link rather than being encoded inline on every request, so responses return faster.

• Chatbot lists, master chat, marketplace, billing pages, and login now do a fraction of the database work they used to, regardless of how many records you have.

Impact

Affects all users, plus every visitor who loads a chat widget. Pages load faster and stay fast as your account grows.

Live Support Option

We've added a new live support option to help you out!

(NEW) 📞 Daily Support Calls Need help getting started or figuring something out? Join our new daily 30-minute calls with the Stammer team. Available Monday to Friday from 11:30am to 12:00pm PST. Link to join: https://stammer.ai/daily-support-call

What's New

Eight new AI models are now available to power your chatbots and agents: the GPT-5.6 family (Sol, Terra, Luna), the Claude 5 family (Sonnet, Opus, Fable), and Grok 4.5 and 4.3.

More Detail

• Select any of the new models from the model dropdown in your chatbot's settings, and make them available to sub-accounts through your agency plans and packages.

• All eight models support image understanding, so users can send images in conversations with any of them.

• Usage is billed through the wallet like any other model, with each model's rates shown clearly in your transaction details.

• GPT-5.6 models manage their response creativity automatically, so the temperature setting is handled for you.

Impact

Affects all Agency Users and Sub-accounts. You can immediately upgrade your agents to the latest generation of AI models for smarter, more capable responses.

(Voice) Add Your Own Custom Voices

What's New

You can now search for and add custom voices from ElevenLabs, Cartesia, MiniMax, and Fish Audio, and use them on your voice agents, no longer limited to the built-in voice list.

More Detail

• Click the new Add voice button beside the voice selector on your voice agent.

• Pick a provider tab, then paste a voice name or the provider's voice ID, each tab links out to that provider's community voice library so you can browse for voices there first.

• The top matching voice appears with its name, age, accent, and gender, plus a Custom badge; click Add to register it.

• The added voice is selected immediately and appears in that account's voice dropdown from then on.

• Custom voices are scoped to your account, other accounts can't see or use the voices you add.

Impact

Affects Agency Users and Sub-accounts using Voice Agents. Give every client a unique, on-brand voice by pulling from thousands of community voices instead of the standard lineup.

Note

Existing plan-based voice restrictions still apply when selecting voices.

Faster, More Resilient AI Responses During Provider Slowdowns

What's Changed

AI requests now have strict time limits and can automatically fail over, so a slow or degraded AI provider no longer drags down the whole platform.

More Detail

• If the AI model takes too long to respond, the request fails fast instead of hanging, keeping the rest of the platform responsive.

• If your agent has a backup model configured, repeated delays or provider failures automatically switch responses to the backup model until the primary recovers.

Impact

Affects all users. During AI provider incidents your chatbots keep answering (via the backup model) and the dashboard stays fast, instead of everything slowing down together.

(Chat) Instagram: Image + Follow-up Text Now Answered Together

What's Changed

When a user sends an image and then a quick follow-up message like "What's this?", the agent now treats them as one question and gives a single answer about the image, instead of replying twice and losing the image context on the second reply.

Impact

Affects all users with agents connected to Instagram. A very common pattern: photo first, question second, now gets one accurate answer instead of a correct reply followed by a confused "I'm not sure what you're referring to."

Note

To merge rapid back-to-back messages, the agent now waits a few seconds after each incoming message before replying, so responses on these channels may arrive after a short pause.

Billing Accuracy & Reliability Improvements

What's Changed

This release includes a set of behind-the-scenes improvements to make billing more accurate and dependable across the platform.

More Detail

• Added safeguards so charges, top-ups, and renewals are always processed correctly and exactly once.

• Improved how subscription and payment statuses are tracked and displayed.

• Made billing operations more reliable, so balances and purchases always reflect what was actually charged.

Impact
Affects all Agency Users and Sub-accounts. Wallet balances, purchases, and subscription statuses are now more accurate and reliable, no action is required on your part.

Billing → Activity → Sub-account Transaction Details Now Show Cost Breakdown

What's Changed

Fixed sub-account wallet transaction details showing "-" for Input Cost, Output Cost, and Total Cost while the matching agency transaction showed the full breakdown. Sub-account transactions now record and display their own complete cost breakdown alongside token counts.

Impact

Affects Sub-accounts and the Agency Users who support them. Sub-account owners can now see exactly what each message cost, without needing to cross-reference the agency view.

Security & Account-Protection Hardening

What's Changed
This release includes a broad security hardening pass across the platform, proactively closing gaps identified in an internal security review.

More Detail

• Strengthened authentication, ownership, and access controls across the platform.

• Hardened internal credential and key management, and updated third-party dependencies to their latest secure versions.

Impact
Affects all users. Your account and data are better protected, no action is required on your part.

Live Support Option

We've added a new live support option to help you out!

(NEW) 📞 Daily Support Calls Need help getting started or figuring something out? Join our new daily 30-minute calls with the Stammer team. Available Monday to Friday from 11:30am–12:00pm PST. Link to join: https://stammer.ai/daily-support-call

What's Changed

As part of our ongoing investment in security, we've rolled out a broad set of enhancements that add extra layers of protection across the platform. Your agency and each sub-account now benefit from strengthened security, privacy, and reliability throughout your account.

More Detail

• We've reinforced protection across the platform to keep your data private and secure.

• We've added extra safeguards and layers of assurance behind the scenes.

• We've strengthened the overall reliability and integrity of your account.

Impact

Affects all Agency Users and Sub-accounts. These enhancements give every account added confidence that its data stays private, secure, and under your control.

Note
These are behind-the-scenes enhancements. Your day-to-day experience is unchanged, there's nothing you need to do.

(Billing) Review Add-ons After a Plan Upgrade

What's New

When you upgrade your plan, the platform now helps you spot add-on subscriptions you may no longer need because your new plan already includes them, and lets you cancel any of them right away.

More Detail

• Right after a plan upgrade, a "Review add-ons" dialog appears if you have existing add-on subscriptions (such as extra storage, extra voice agents, extra concurrency, extra voice knowledge base, or extra chat agents) that your new plan may now cover.

• Each add-on is listed with its quantity, amount, and next recharge date so you can decide with full context.

• Tick the add-ons you want to drop and choose "Cancel Selected Add-ons" to schedule their cancellation, or choose "Keep My Add-ons" to close the dialog and change nothing.

• You'll get a confirmation once the selected cancellations are scheduled; anything you don't select stays active as before.

Impact

Affects Agency Users (and any account with add-on subscriptions) who upgrade their plan. It helps you avoid paying for extras your new plan already includes, giving you a clear moment to trim redundant costs.

Note
The prompt only appears when a relevant add-on exists after an upgrade, and cancellations are scheduled, selected add-ons remain in effect until their scheduled cancellation takes place.

Live Support Option

We've added a new live support option to help you out!

(NEW) 📞 Daily Support Calls

Need help getting started or figuring something out? Join our new daily 30-minute support calls with the Stammer team.

• Schedule: Monday to Friday, 11:30 AM – 12:00 PM PST

• Join here: https://stammer.ai/daily-support-call

What’s New

You can now generate a complete base system prompt for an AI agent just by describing what you want in plain language.

The builder writes a ready-to-use prompt from your description and lets you refine it in one click.

More Detail

• Open an agent’s Base System Prompt during agent creation or while editing an existing agent, then launch the AI Prompt Builder.

• Describe your business and the agent you want, including its role, tone, and what it should help customers with, then click Generate.

• Once a draft appears, refine it with one click:

  • Regenerate for a fresh take

  • Improve for a clearer and more complete version

  • Make It Shorter

  • Make It Sales-Focused

• The builder uses clear placeholders in square brackets, such as [insert booking link], instead of inventing facts.

• It automatically appends the standard default guideline lines for you to edit or remove.

• Review the draft, make any necessary changes, and save it to your agent.

Impact

Affects all Agency Users and Sub-accounts who create or edit AI agents.

It dramatically reduces the effort required to write a high-quality prompt and helps new users launch a well-structured agent faster.

Note

The builder only returns text for you to review and does not update your agent until you save it.

Access follows the same permissions as editing an agent, and generation is rate-limited to keep things responsive.

💬 Inbound Media on WhatsApp & Instagram

What’s New

Your agents can now receive media sent by customers including images, documents, audio, and video, instead of ignoring anything that is not plain text.

More Detail

• When a customer sends media through a connected WhatsApp or Instagram account, the file is captured, stored, and shown in the conversation.

• The agent can take the incoming media into account without dropping the message.

• Outgoing agent replies that include an image are now delivered correctly over WhatsApp instead of failing due to unsupported content.

Impact

Affects Agency Users and Sub-accounts using WhatsApp and Instagram.

Conversations feel more natural and complete because customers can share pictures and files, and your agent no longer goes silent on non-text messages.

Note

This is Part 1 of a phased rollout of media handling. Additional improvements will follow in later releases.

Hardened Access Controls for User & Permission Management

What’s Changed

We strengthened the controls around managing users and their permissions in agency and sub-account accounts, ensuring permission changes are limited to the appropriate administrators.

More Detail

• Only agency admins can manage agency users.

• Only sub-account admins can manage sub-account users.

• Users can no longer change their own role or permissions.

• Permission assignment is now restricted to a safe, approved set of options.

Impact

Affects all Agency Users and Sub-accounts.

These tighter controls ensure sensitive capabilities can only be granted by an authorized administrator.

💬 Improved Context Retention in Long Conversations

What’s Changed

Agents now retain context more reliably during long conversations, helping them remain consistent instead of “forgetting” earlier parts of the conversation.

More Detail

• The running conversation summary that keeps your agent informed is now generated and reused correctly as chats grow longer.

• This reduces cases where an agent loses track of details shared earlier in the same conversation.

Impact

Affects all users running chat agents, especially those with longer conversations.

Customers receive more coherent and relevant answers throughout the entire chat.

My Chat Agents Page: Faster Loading & Unified Skeletons

What’s Changed

The My Chat Agents page now loads noticeably faster and displays improved placeholders while fetching your agents.

More Detail

• Agent and message counts are now calculated more efficiently, speeding up the page—especially for accounts with many agents and large conversation histories.

• Loading states across the page now use unified skeleton placeholders for a smoother and more polished experience while content loads.

Impact

Affects all Agency Users and Sub-accounts.

Users will experience less waiting and a smoother experience when reviewing chat agents.

Wallet Activity: Accurate Per-Row Balances

What’s Changed

The balance shown on each row of the Wallet Activity tab now correctly reflects your actual wallet balance at the time of that transaction.

More Detail

• Reconciled transactions now record the running balance in the same way as all other wallet activity.

• Each row’s balance now aligns with your actual wallet balance.

• This resolves cases where a row’s displayed balance did not match the real balance.

Impact

Affects all wallet users.

Your wallet history is now more accurate and reliable when reviewing past transactions.

Sub-Accounts → Billing Status: Clearer Update Messages for Stripe-Billed Accounts

What’s Changed

When you attempt to manually change the billing status of a sub-account billed through our payment provider, you now receive a clear explanation of why the change is not allowed.

More Detail

• Manual billing-status overrides are blocked for sub-accounts with an active paid subscription.

• The reason for the restriction is now clearly stated.

• Related validation messages, including trial-end messages, have also been improved for clarity.

Impact

Affects Agency Users managing sub-account billing.

This reduces guesswork and prevents surprises when adjusting how a sub-account is billed.

Behind the Scenes: Leaner Unread-Status Storage

What’s Changed

We optimized how conversation read and unread statuses are stored, significantly reducing database usage as conversation volume grows.

More Detail

• Redundant read-status records are cleaned up.

• Read-status data is stored using a more efficient index.

• Unread indicators remain fast and reliable at scale, with no change to how they work for users.

Impact

Affects all users, especially high-volume accounts.

This improves the performance and reliability of unread indicators, with no action required on your part.

Live Support Option

We’ve added a new live support option to help you!

NEW: 📞 Daily Support Calls

Need help getting started or figuring something out?

Join daily support calls with the Stammer team, available Monday through Friday from 11:30 AM to 12:00 PM PST.

Link to join:
https://stammer.ai/daily-support-call

What's New

Your chat and voice agents can now book Calendly appointments that take place over the phone, including outbound calls where the agent collects the customer's callback number during the conversation.

More Detail

• In your agent's Calendly scheduling setup, the location list now includes phone-based options (inbound call and outbound call) alongside the existing in-person/physical location option, pulled directly from your selected Calendly event type.

• When you choose the outbound call option, the agent automatically asks the customer for the phone number they'd like to be called on, confirms it before booking, and saves it with the appointment—no manual configuration needed.

• The sample prompt shown in the scheduling settings updates automatically to include the outbound-call instructions, so you can copy a ready-to-use prompt for your agent.

• Booking error messages are now more specific when Calendly declines a booking (for example, if a slot is no longer available), making it easier to understand what went wrong.

Inbound:

Outbound:

Dashboard:

Impact

Affects Agency Users and Sub-accounts using Calendly scheduling on chat or voice agents. You can now offer phone-call appointments end-to-end, letting your agent capture a callback number during the conversation instead of forcing customers into video or in-person meetings.

Note

Available for event types configured with phone-based locations in Calendly. Existing scheduling setups remain unchanged unless you switch to a phone-based location.

Conversation Export CSV: Per-Message Timestamps & Activity-Based Filtering

What's New

Conversation CSV exports now include the date and time of each individual message, plus smarter date filtering that captures conversations with recent activity.

More Detail

• Exported CSVs now include a Message Date/Time column on each message row, so you can see exactly when every user message and agent response occurred.

• A Conversation Created Date/Time value is included in the conversation details, separate from the per-message timestamps.

• Date range filters now use activity-based filtering, including conversations created within the selected range as well as conversations that received new messages during that period.

Impact

Affects all users who export conversation data. Reporting and analysis are more accurate with message-level timestamps and activity-based filtering.

Faster Conversation Reset

What's Changed

Resetting a conversation now feels nearly instant, with heavier cleanup tasks moved to the background.

More Detail

• When you reset a conversation, the action completes immediately instead of waiting for follow-up processing (such as lead analysis and conversation summaries), which now runs asynchronously.

Impact

Affects all users who reset conversations in the chat widget. Reset actions complete much faster, allowing you to continue without waiting for background processing.

(Chat) Message Rate Settings: Clear Error for Long Cooldown Messages

What's Changed

Saving a Message Rate cooldown message longer than 255 characters now displays a clear validation error instead of silently failing.

More Detail

• The cooldown message field is validated at 255 characters.

• If exceeded, the field displays: "Cooldown message must be 255 characters or less."

• Server validation errors are now mapped back to the correct fields in the Message Rate settings form.

Impact

Affects users configuring Message Rate limiting on chat agents. Validation errors are now clear and actionable.

Conversation Filters Now Apply Correctly in Inbox

What's Changed

Fixed an issue where Inbox conversation filters were not being applied correctly.

Impact

Affects all users browsing conversations in the Inbox. Filters now reliably narrow results as expected.

Custom Functions: Parameter Name Validation

What's Changed

Custom Functions now validate parameter names and prevent saving names that contain spaces.

More Detail

• Parameter names (dynamic or static) containing spaces are blocked during save.

• A clear validation message explains that parameter names cannot contain spaces.

• This prevents malformed parameters that could cause runtime failures.

Impact

Affects Agency Users and Sub-accounts building Custom Functions. Invalid parameter names are caught before deployment.

White-Label Embed: Removed Platform-Identifying References

What's Changed

Public chat widget embed code no longer contains platform-identifying references, keeping white-label implementations fully branded.

More Detail

• The embed loader now determines its backend domain from your configured domain (or the domain serving the script) instead of relying on platform-specific attributes or hardcoded fallback domains.

Impact

Affects Agency Users using white-label chat widgets on their own or client websites. Embedded widgets remain fully branded to your domain.

Note

Existing embeds using a configured domain continue to work without changes.

Live Support Option

We've added a new live support option to help you out!

(NEW) 📞 Daily Support Calls

Need help getting started or figuring something out? Join our new daily 30-minute calls with the Stammer team.

Schedule: Monday–Friday, 11:30 AM–12:00 PM PST

Link to join: https://stammer.ai/daily-support-call

What's New

You can now choose whether to be notified about partial leads when a conversation ends. A new toggle on the chat agent's lead notification settings lets you receive an alert as soon as a conversation wraps up, even if the visitor didn't fill in every lead field.

More Detail

- When enabled, you get notified even if not all configured lead fields were collected.

- When disabled, notifications are only sent once a lead is fully completed (all fields captured).

Impact

Affects all users collecting leads through chat agents. Make sure no promising-but-incomplete lead slips through, or keep notifications focused only on fully qualified leads.

Note

Enabled by default, so you'll start receiving partial-lead notifications unless you switch this off per chat agent.

Sub-Accounts → New Table Design with Date-Range Comparison

What's New

The sub-accounts dashboard has a redesigned table that shows spend per sub-account for a date range you choose, and lets you compare it against a second date range side by side.

More Detail

- Pick a primary date range to see each sub-account's spend over that period.

- Add a comparison date range to view performance against a previous period at a glance.

Impact

Affects Agency Users managing sub-accounts. Track sub-account activity and spending trends over time and spot changes between periods without exporting data.

Sub-Account Alert Rules

What's New

Agency owners can now create alert rules from the sub-accounts dashboard and automatically receive an email whenever a sub-account crosses a threshold you define. Set up the conditions once, and Stammer watches your sub-accounts for you.

More Detail

- Name each rule so it's easy to recognise later, for example, "Wallet balance below $10."

- Choose what to monitor:

- Wallet balance: track a sub-account's remaining balance.

- Usage amount: track spend over a rolling window of the last 1, 7, or 30 days.

- Set the condition: fire when the value goes above or below your dollar threshold.

- Target the rule: apply it to all sub-accounts at once, or to a specific sub-account.

- Enable or disable any rule without deleting it, so you can pause monitoring and turn it back on later.

- Get a clear email when a rule triggers. The notification names the sub-account, the rule that fired, the exact condition (e.g. "wallet balance below $10"), and the current value at the moment the threshold was crossed — so you know what happened at a glance and can act right away.

Impact

Affects Agency Users managing sub-accounts. Stay ahead of low wallet balances or unexpected usage spikes without manually checking each sub-account, and keep a closer eye on the accounts that matter most.

Note

Alert emails are sent to the agency admin's email address. New rules are enabled by default.

(Voice) Correct Toast When Un-assigning a Phone Number

What's Changed

Un-assigning a phone number from a voice agent now shows the correct "Phone number unassigned successfully" message instead of incorrectly saying it was assigned. Error messages were corrected to match the action as well.

Impact

Affects Agency Users and Sub-accounts managing voice agents. Clear confirmation that matches the action you just took.

Instagram Integration Now Shows the Connected Account Name

What's Changed

The Instagram integration page now displays the connected account name. When a profile's display name isn't available, it falls back to showing the username instead of appearing blank.

Impact

Affects users connecting Instagram to their chat agents. Easily confirm which Instagram account is connected.

(Chat) Conversation Summary Emails Now Accept Long Domain Endings

What's Changed

Email validation for conversation summary recipients no longer rejects valid addresses that use longer domain endings (for example, addresses ending in .travel). Email entry was also made more robust against empty values.

Impact

Affects users sending conversation summary emails. Add any valid recipient address without running into false "invalid email" errors.

Hide Embed Code When User Lacks Embed Permission

What's Changed

The embed code section in the chat agent detail side panel is now hidden for users who don't have the Embed Chatbot permission.

Impact

Affects Agency Users and Sub-accounts with restricted permissions. Keeps the interface aligned with each user's permissions and prevents embedding by users who shouldn't have that access.

Cap on Concurrent Messaging API Connections per IP

What's Changed

We added a safeguard that limits how many simultaneous Messaging API connections a single source can hold at once.

Impact

Affects all users relying on chat and the Messaging API. Protects platform stability and responsiveness by preventing any single source from overloading the service.

Note

This is an abuse-prevention safeguard and does not affect normal usage.

---

Live Support Option

We've added a new live support option to help you out!

📞 Daily Support Calls

Need help getting started or figuring something out? Join our new daily 30-minute calls with the Stammer team. Available Monday to Friday from 11:30am–12:00pm PST.

Link to join: https://stammer.ai/daily-support-call

What’s New
Chatbots can now be shared through links that include an expiration setting, allowing teams to grant temporary access without leaving links active indefinitely.

More Detail

• Expiration-based sharing helps agencies control external access windows (for demos, client reviews, or limited-time testing).

• Once a link expires, access through that URL is automatically blocked.

• Reduces risk from old links being reused or circulated after they are no longer intended to be active.

• Improves operational control when multiple team members are sharing bots with prospects or clients.

Impact
Affects agency users and sub-accounts sharing bots externally. Improves security posture, access governance, and confidence when sharing chatbot experiences outside the platform.

Note

All previously generated share links are not affected by this update and will continue to behave as they did before.

Billing → Usage & Settings → Payments Showing Receipts Instead of Invoices

What’s Changed
Corrected the billing display behaviour in Billing → Usage & Settings → Payments so users see the intended financial document context more clearly.

Impact
Affects all users reviewing billing history. Reduces confusion in payment records and improves reconciliation workflows.

Wallet Restructuring & Locking Mechanism for Concurrent Recharges

What’s Changed
Wallet processing has been reworked to introduce stricter transaction handling with a locking mechanism during recharge operations, preventing race conditions caused by simultaneous requests.

More Detail

• Recharge flows now enforce controlled write access during balance updates, so overlapping requests cannot mutate the wallet state at the same time.

• Prevents common concurrency failures such as duplicate credits, partial state updates, or inconsistent post-recharge balances.

• Improves idempotency behaviour around near-simultaneous recharge attempts from retries, rapid user clicks, or parallel system triggers.

• Strengthens ledger consistency between recharge events and resulting wallet balances.

• Reduces need for manual correction in edge cases where concurrent operations previously created mismatch risk.

Impact
Affects all wallet users. Significantly improves financial consistency, recharge reliability, and trust in wallet balance accuracy under high-concurrency conditions.

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

The entire platform has been rebuilt with a unified React design system, consistent card surfaces, tab patterns, typography, and spacing applied across every page. Previously, every page navigation triggered a full server round-trip and reload. Now, only data is fetched, pages switch instantly, common routes are prefetched in the background, and the browser only loads code for the section you're using. Switching themes no longer requires a reload. The result is a noticeably faster, smoother experience across the entire platform.

What's New

Sidebar Overhaul

The sidebar has been fully rebuilt. Navigation is now cleaner and more organised, with a collapsible design that supports both expanded and icon-only modes.

• Account Settings and Billing moved to bottom-left user profile panel

User panel includes:

• Account Settings

• Billing

• Theme toggle

• Language selector

• Logout

New Account Stats widget above user panel showing:

• Wallet balance

• Chat Agents

• Voice Agents

• KB Slots

• Concurrency

• Character usage

• Inline expandable stats without leaving the page

• Sub-account switching via inline searchable popover (no page reload)

Tab Relocations Across the App

The Vault and Users pages, previously standalone sections in the navigation, have been consolidated into Account Settings as dedicated tabs.

Account Settings now contains:

• Profile

• Notifications

• Custom Menu Links

• Default Prompts

• Email Templates

• Vault

• Users

• Analytics tab moved to the last position in the AI Agent tab bar

• Subscription and Upgrade Plan moved into the Billing tab of the Billing dashboard

Dashboard Redesign

The main dashboard now uses an 8/4 column grid layout with synced card heights and primary-coloured section headings.

• Sticky cards and headers keep context visible while scrolling

Unified Card & Tab Patterns

All dashboards now share a consistent card-based layout with standardised tab styling.

Applies to:

• Voice

• Knowledge Base

• Integrations

• Appearance

• Marketplace

• Account Settings

• Chatbot

• Agency Settings

Tables now include:

• Consistent row spacing

• Unified colours and borders

• Empty states with inner scroll

• Buttons, chips, status badges, and indicators are fully theme-driven

• Filter dropdown redesigned as a popover with active-filter indicator

Impact

Affects all users on all pages. No configuration changes required, purely a structural, navigational, and visual improvement.

Faster Navigation: Route Skeletons & Prefetching

Page transitions are now smoother and loading states are more informative.

What's New

Every page now has a purpose-built loading skeleton that matches the shape of the loaded content. Spinners have been replaced with structured skeleton screens across the dashboard, chatbot, onboarding, auth, and marketplace pages.

Common sidebar routes are prefetched in the background so navigating between sections feels instant.

(Voice) Normalize for Speech

Voice agents can now automatically convert numbers, dates, currencies, and similar text into a natural speech-friendly format before speaking them aloud.

What's New

A new Normalize for Speech toggle has been added to the Voice Agent call settings. When enabled, values like "£1,200.50" or "2026-04-22" are converted into how a person would naturally say them before being passed to the voice.

Note: enabling this option introduces a slight processing delay before the agent responds.

Impact

Affects all Agency Users and Sub-accounts using Voice Agents. Disabled by default, opt in per agent.

(Voice) Dynamic Speed & Responsiveness

Voice agents now automatically adapt their speaking speed and responsiveness to match the pace and tone of the conversation.

What's New

Dynamic voice speed and dynamic responsiveness are now enabled across all voice agents. The agent naturally adjusts how quickly it speaks and how promptly it responds based on the flow of the conversation, without any manual configuration.

Impact

Affects all Agency Users and Sub-accounts using Voice Agents. Applied automatically, no configuration required.

(Voice) Post-Call Analysis Model Updated

The model used to process post-call analysis including lead extraction and conversation summaries has been updated to GPT-4.1 Mini.

What's Changed

Post-call analysis now runs on GPT-4.1 Mini, a faster and more capable model than its predecessor.

Impact

Affects all Voice Agent users with post-call analysis enabled. Lead extraction and summaries will be more accurate and processed faster.

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

The billing dashboard has been fully rebuilt in React with a clean three-tab layout covering subscriptions, transaction history, and add-on management.

What's New

The dashboard is organised into three main tabs: Billing, Activity, and Add-Ons.

• Billing Tab: View your current plan and status, manage saved payment methods, and configure wallet auto-billing thresholds. The Subscription / Upgrade Plan section has been moved here from the sidebar, centralising all billing actions in one place.

• Activity Tab: Full transaction history with four sub-tabs: All, Wallet, Storage, and Payments. Each supports search, date filtering, CSV export, and a details modal per row.

  • Payments contains two sub-tabs: Payments (charges made by your account) and Income (incoming payments, visible to agency accounts only). Invoices are shown when available, with receipts as a fallback.

• Add-Ons Tab: Contains two sub-tabs: Purchase Add-Ons (extra Storage, Chat Agents, Voice Agents, Concurrency, or Voice KB Slots) and Active Add-Ons (manage existing add-on subscriptions with per-add-on auto-recharge settings).

Impact

Affects all Agency Admins and Sub-account Users. Centralises subscription, transaction, and add-on management into a single, consistent interface.

(Chat) Human Handoff: Intent Recognition & Dynamic Transfer to Human Agent

Users can now configure how and when their AI agent hands off a conversation to a live human agent, choosing between keyword-based or AI intent-based triggering.

What's New

• Intent Mode: The AI automatically detects when a user wants to speak with a human and triggers handoff, no keyword required. To enable this, copy the provided intent prompt and paste it into your agent's Base Prompt, this is required for intent detection to function correctly.

• Keyword Mode: Handoff triggers when a user message matches a configured keyword list (e.g., "Talk to human", "STOP").

• Handoff notifications (email/web-hook) fire immediately when a transfer is triggered.

• Supported across Chat Widget, WhatsApp, Instagram, Facebook Messenger, and API channels.

• Human Handoff settings have been moved to the Settings tab of the AI Agent for easier access and a cleaner configuration layout.

Impact

Applies to all Agency Users and Sub-accounts using the Chat Agent with Human Handoff enabled. Improves handoff reliability and reduces missed escalation scenarios.

New AI Agent Dashboard Experience

The AI Agent configuration dashboard has been redesigned for clarity and ease of use.

What's New

• Separate tabs for AI Agent, Master Chat, and Settings configuration.

• Consolidated agent usage stats displayed in summary cards.

• Cleaner layout with improved information hierarchy.

• Faster access to frequently used settings without scrolling.

Impact

Affects all users managing AI Agents. The new layout reduces cognitive load and makes configuration faster to navigate.

Revamp Vault Dashboard

The Vault Dashboard has been fully rebuilt in React, replacing the legacy template-based UI.

What's New

• Tab-based navigation separates vaults by LLM provider: OpenAI, Anthropic, X.AI, and ElevenLabs. Note: X.AI links directly to the provider's API key page; the other providers do not currently include a direct link.

• Skeleton loading states for improved perceived performance.

• New custom icons and Material-UI responsive layout.

• Full multi-language support added (300+ translation keys across all supported locales).

Security Update

API keys entered in the Vault are now transferred between the frontend and backend using end-to-end encryption, preventing man-in-the-middle attacks during key submission.

Impact

Affects Agency Admins managing LLM API keys via the Vault. The new UI is more intuitive, consistent with the rest of the platform, and more secure.

Show "No Leads Collected" Message in Summary Emails When No Leads Exist

Conversation summary emails now handle missing and empty lead data explicitly.

What's Changed

• When no leads were collected during a conversation, summary emails now display: "No leads were collected in this conversation."

• If a lead exists but an optional field was not filled in, that field now displays "Not specified" rather than appearing blank.

• Translated into all supported languages.

Impact

Applies to all chatbot owners receiving conversation summary emails. Eliminates ambiguity around both missing leads and incomplete lead entries.

Billing, Usage & Settings → Activity: Invoices Now Shown Correctly

The Payments tab was incorrectly displaying receipts in place of invoices.

What's Changed

• The table now correctly prioritises invoices when available.

• Falls back to receipt only when no invoice exists.

• The document type label dynamically reads "Invoice" or "Receipt" based on what is available.

• Column header updated to Invoice / Receipt for clarity.

Impact

Affects all users viewing payment history in the Billing section. Ensures compliance-relevant documents (invoices) are surfaced correctly.

Add Scrollable Input Field + Collapsible Long AI Responses ("Read more")

Two usability improvements to the chat widget for both users and agents.

What's New

• Scrollable Input: The message input field now scrolls vertically when the message exceeds the visible area, preventing the input from growing unbounded.

• Collapsible AI Responses: Long AI responses are now truncated with a "Read more" toggle. Users can expand or collapse the response as needed.

Impact

Affects end users interacting via the Chat Widget. Improves readability and reduces layout disruption from lengthy AI responses.

Standardise "View History" Button Styling Across the App

Inconsistent styling for "View History" buttons has been resolved across the platform.

What's Changed

• A unified View History button style is now applied consistently across Custom Functions, Email History, and Web-hook History pages.

• Consistent styling, alignment, and behaviour regardless of where the button appears.

Impact

Visual consistency improvement affecting all users navigating history views. No behavioural changes.

LLM Model Dropdown Menu UI Update

The model selection dropdown has been redesigned for better usability.

What's Changed

• The dropdown is now scrollable when many models are available.

• Improved styling and visual clarity.

• Applies to both the primary LLM model and the confirmation model selectors.

Impact

Affects all users configuring AI Agent settings. Particularly helpful for accounts with access to a large number of LLM models.

Updated Incorrect German Translations in Dashboard & Subscription Pages

A comprehensive pass was made over German locale strings across the platform.

What's Changed

• 365+ German translation keys corrected across Dashboard and Subscription pages.

• Fixes include Human Handoff terminology, AI Agent configuration labels, and billing UI strings.

• All 13 supported language files reviewed for consistency.

Impact

Affects all users operating the platform in German. Incorrect or untranslated strings have been replaced with accurate German equivalents.

(Chat) Notify Agent Owner When Agent Cannot Answer (KB Gap Detection)

Agents can now alert their owners when they encounter questions they cannot answer.

What's New

• New toggle: Notify when agent cannot answer, sends an email to the agent owner when a knowledge base gap is detected.

• Additional notification recipients can be configured.

• Optional web-hook notification can be enabled with a custom URL.

• A 60-minute deduplication window prevents repeated notifications for the same unresolved topic.

• Applies to both Chat and Voice agents.

Impact

Affects Agency Users and Sub-accounts using Chat or Voice agents. Enables proactive knowledge base improvement by surfacing unanswered questions in real time.

(Chat UI) Improve Image Rendering Quality & Update Conversation Starter Layout

Visual polish improvements to the embeddable Chat Widget.

What's Changed

• Image rendering quality improved for sharper display in conversation threads.

• Conversation Starter suggestions have an updated layout for better visual presentation and responsiveness.

Impact

Affects end users interacting with Chat Widgets. No configuration changes required.

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

The Subscription tab has been merged into Billing Usage & Settings, along with significant UI/UX improvements to create a more unified and streamlined billing experience.

What’s New

Unified Billing & Subscription Management

All subscription-related functionality is now fully integrated in the Billing Usage & Settings section, eliminating the need for a separate Subscription tab.

Improved Layout & Structure

The billing interface has been reorganised into clear sections, including:

• Plan & subscription overview

• Payment methods

• Agency wallet & auto-billing

• Usage activity (wallet, storage, payments, income)

• Add-ons and resource allocations

Enhanced Visibility & Controls

Users can now manage:

• Subscription plans

• Payment methods

• Wallet balance and auto-recharge

• Add-ons (storage, agents, concurrency, etc.)

—all from a single, centralised interface.

Impact

• Centralised billing, usage, and subscription management

• Improved clarity across billing data and controls

• Faster navigation and reduced friction when managing plans and resources

• More scalable structure for future billing features

Embedded Voice Agent Widget

Voice Agents can now be embedded directly on websites using a new widget-based voice experience, allowing users to test and use voice agents without requiring a phone number.

What’s New

Embeddable Voice Widget

Users can now assign a Voice Agent to a website widget and embed it using either:

• Iframe embed code

• Script embed code

This makes it possible to launch a Voice Agent directly on a website through a clickable voice orb.

Customisable Widget Appearance

The widget includes configuration options for:

• Orb style

• Colour theme

• Primary and secondary colours

• Widget size

• Widget position

• Custom logo upload

Live Preview

A live preview panel allows users to see how the widget will look and interact before saving.

Separate Widget Assignment

Voice Agents can now be configured for widget use independently, allowing the same agent behaviour to be used through the widget experience.

Impact

• Allows users to test and use Voice Agents without purchasing a phone number

• Expands Voice Agent deployment options to websites

• Provides flexible customisation for branding and appearance

• Makes it easier to use the same Voice Agent across different channels

Voice AI Agent Billing Updates

Voice AI agent billing has been updated with clearer pricing and improved control over add-ons.

Reduced Pricing

• Additional Voice Agents: $5/month per agent

• Voice Knowledge Base (KB): $10/month for 1 KB slot (≈5M characters)

• Phone Numbers (USA/CAN): $2/month per number

• Concurrent Calls: $10/month per concurrent call

Updated Voice Agent Quotas

Now,

• Agency Plan: Includes, 20 Voice Agents

• Full SaaS Mode: Includes, 100 Voice Agents

• Enterprise: Includes, 250 Voice Agents

• Enterprise Level 2: Includes, 100 Voice Agents

Note: All existing add-on subscriptions have been discontinued, and the new pricing takes effect immediately.

What’s New

Voice Agent Allocation in Plans
Subscription plans now clearly show the number of included Voice Agents, making it easier to understand capacity at each tier.

Add-On Billing Improvements
Voice-related add-ons have been streamlined and are now configurable directly in Billing Usage & Settings → Add Ons, including:

• Extra Voice Agents

• Extra Voice KB Slots

• Extra Concurrency

Each add-on includes structured pricing tiers and auto-billing options.

Impact

• Transparent and predictable voice pricing

• Easier scaling with clear per-unit costs

• Better control over auto-recharge and usage thresholds

• Improved visibility of voice resources across plans and add-ons

Configurable Timeout for Custom Function API Calls

Timeout configuration is now available for custom function API calls, giving users more control while maintaining system stability.

What’s New

Custom Timeout Field

Users can now define a timeout (in seconds) directly when creating or editing a custom function. This allows better handling of APIs with varying response times.

Upper Limit Enforcement

To prevent failures and degraded performance, a maximum timeout limit is enforced. Users cannot exceed this predefined cap.

Warning Messaging

A clear warning is displayed when increasing the timeout, helping users understand potential risks such as slower responses or failed calls from the widget.

Input Validation

Timeout values are validated to ensure they remain within acceptable limits.

Impact

• Greater flexibility for integrating external APIs

• Reduced risk of broken or hanging function calls

• Improved reliability of widget and agent interactions

• Better user awareness when configuring advanced settings

WhatsApp Connection Disabled

WhatsApp connection has been disabled in cases where message delivery fails due to configuration or verification issues.

What’s Changed

Automatic Disabling on Errors

The system now automatically disables WhatsApp connections when critical errors occur, such as:

• Phone number not verified

• Invalid or incomplete setup

• Other delivery-related failures

Impact

• Prevents repeated failed message attempts

• Improves system reliability and error handling

• Ensures users are aware of configuration issues before retrying

Users can re-enable the connection once the underlying issue (e.g., number verification) is resolved.

Cal.com Region Selector & Dynamic API Base URL

The Cal.com integration now supports a region selector and uses a dynamic API base URL based on the selected region.

What’s New

Region Selector in Cal.com Settings

A new region option has been added to Cal.com settings with the following choices:

• Global

• EU

Dynamic API Base URL

API requests now use the correct base URL depending on the selected region:

• Global → https://api.cal.com/v2

• EU → https://api.cal.eu/v2

Default Behavior for Existing Integrations

Existing Cal.com integrations default to Global.

Impact

• Ensures EU-hosted Cal.com customers use the correct API endpoint

• Prevents API failures caused by using the wrong base URL

• Improves compatibility for both Global and EU Cal.com users

Custom Menu Links Revamp

Custom Menu Links have been revamped and moved to Agency → White Label for a more centralised white-label configuration experience.

What’s New

New Location in White Label Settings

Custom Menu Links are now managed from the Custom Menu Links tab under Agency → White Label.

Improved Management Experience

The updated interface provides a cleaner flow for creating and managing custom navigation links for sub-accounts.

Users can configure:

• Icon

• Label

• URL

• Open in new tab option

Impact

• Centralises custom menu link management within white-label settings

• Improves discoverability and usability

• Makes it easier to manage custom sidebar navigation for sub-accounts

Consent Checkbox for White-label

A consent checkbox has been added to the White-label Create Account page.

What’s New

Users must now confirm they agree to the Terms of Service and Privacy Policy before creating an account.

The checkbox uses the Privacy Policy URL and Terms of Service URL configured in White Label Settings → Dashboard Settings.

Impact

• Improves compliance

• Adds clear user acknowledgment at sign-up

• Keeps policy links configurable from the dashboard

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

Chat Agent Integration with Calendly & Cal.com

Chat Agents now support Calendly and Cal.com integrations for appointment scheduling, replacing the previous Google Scheduling system.

Previously, scheduling was only available through Google Scheduling. This legacy method has now been deprecated in favour of more flexible and scalable scheduling integrations.

What’s New

Calendly Integration

Chat Agents can now schedule meetings directly through Calendly, enabling users to book appointments that automatically generate Zoom or Google Meet meetings.

Configuration includes:

• Calendly API Token

• Event Type selection

• Location configuration

• Automated scheduling prompt template

Cal.com Integration

Support has also been added for Cal.com scheduling, providing an alternative scheduling platform with similar capabilities.

Configuration includes:

• Cal.com API Key

• Event Type selection

• Timezone configuration

• Automated scheduling prompt template

Integration Rules & Setup

To ensure proper functionality, the following rules apply:

Legacy Google Scheduling Deprecation

• The previous Google Scheduling integration is now deprecated.

• Users are strongly encouraged to migrate to Calendly or Cal.com for future scheduling workflows.

Legacy Scheduling Conflict

• If Legacy Google Scheduling is enabled, the new integrations cannot be used.

• Legacy scheduling must be disabled before enabling Calendly or Cal.com.

Mutual Exclusivity

Only one scheduling integration can be active at a time:

• If Calendly is enabled, Cal.com cannot be used.

• If Cal.com is enabled, Calendly cannot be used.

This prevents conflicts in the agent’s scheduling logic.

Prompt Integration Requirement

Each scheduler includes a pre-built scheduling prompt template.

To activate scheduling behaviour:

1. Copy the prompt provided in the Tools → Scheduling section.

2. Paste it into the Chat Agent’s Base Prompt.

3. Adjust the instructions if needed to match your scheduling workflow.

Without adding the prompt to the agent’s base prompt, the scheduling tools will not function correctly.

Impact

• Replaces legacy Google scheduling with modern scheduling integrations

• Provides greater flexibility with multiple scheduling providers

• Enables more reliable automated appointment booking workflows

• Improves maintainability and scalability of scheduling features across Chat Agents.

Prompt-Based Dynamic Call Routing

Voice Agents now support prompt-based dynamic call routing, allowing calls to be transferred to different phone numbers based on what the caller says during the conversation.

Previously, calls were routed to a single static phone number. With this update, the agent can now use a prompt-based approach (similar to the dynamic routing option available in the UI) to determine the correct destination number at runtime.

What’s New

Prompt-Based Routing

Instead of routing calls to a fixed number, the Voice Agent now analyses the caller’s intent and determines where the call should be transferred.

Intent-Based Call Transfers

Calls can now be routed to different phone numbers depending on the caller's request.

Example scenarios:

• If the caller says they want support, transfer to: +1 (925) 222-2222

• If the caller says they want sales, transfer to: +1 (925) 333-3333

Support for New GPT & Grok Models

Support has been added for the latest GPT and Grok models to ensure the platform stays up-to-date and users can leverage the newest capabilities.

Supported Grok Models

• Grok-3-mini

• Grok-4-0709

• Grok-4-fast-reasoning

• Grok-4-fast-non-reasoning

Supported GPT Models

• GPT-5.2-2025-12-11

What’s New

Expanded Model Compatibility

Added compatibility with newly released GPT and Grok models for both Chat and Voice Agents.

Automatic Migration Support

Deprecated Claude models are automatically migrated to supported newer models.

Assign Specific Chat Agent Permissions Per User

User permission management has been enhanced to allow assigning access to specific Chat Agents on a per-user basis.

Previously, when creating or editing a user, administrators could only grant general Chat Agent permissions (such as creating chatbots, deleting chatbots, or viewing conversations). Any user with Chat Agent permissions automatically had access to all Chat Agents, with no option to restrict access to selected agents.

What’s New

Chat Agent Assignment in User Management

A new Assigned Chat Agents section has been added during User Creation and User Editing, allowing administrators to control which Chat Agents a user can access.

Granular Access Control

Admins can now:

• Select one or multiple specific Chat Agents a user can access

• Restrict visibility and editing rights to only those assigned agents

Flexible Agent Selection

The Chat Agent assignment interface supports:

• Assign All option

• Individual agent selection

• Search functionality for large agent lists

• Assigned-only view filter

Default Behaviour

• If no Chat Agents are assigned, the user will not have access to any Chat Agents

• Admin users can optionally be granted access to all agents

Impact

• Improves granular access control for agencies managing multiple agents

• Enables restricted access for internal team members, contractors, or client-specific users

• Enhances security and operational control when managing multiple Chat Agents across teams or clients

Chat Widget Improvements & Templates

Several UX improvements have been made to the Chat Widget along with the introduction of widget design templates.

What’s New

Widget Templates

A new Design Templates section has been added to the Appearance settings, allowing users to quickly apply predefined widget styles.

Available templates include:

• Classic Blue

• Dark

• Light

• Midnight Purple

• Classic Teal

• Purple Gradient

• Minimal Clean

These templates allow users to quickly customise the widget’s appearance without manually adjusting individual styling options.

Input Field Auto-Focus Adjustment

The chat input field will no longer automatically receive focus when the iframe loads.
Users must now manually click into the input field before typing.

Multi-line Input Field Expansion

For longer messages, the input field now expands vertically so the entire message remains visible while typing.

Previously, the input field only displayed the current line, making it difficult to review or edit longer messages.

Multi-line Welcome Message Support

Welcome messages now support multiple lines within a single message, preserving line breaks instead of splitting each line into separate messages.

Impact

• Faster widget styling using predefined templates

• Improved typing experience for longer messages

• Better formatting support for welcome messages

• Improved user experience when chat widgets are embedded in iframes

Privacy Policy & Terms Acknowledgement for White-label Sign-Ups

White-label sign-up pages now support compliance confirmation fields.

What’s New

Mandatory Consent Checkbox

A new checkbox allows users to acknowledge:

• Privacy Policy

• Terms & Conditions

during the sign-up process.

Impact

• Improved legal compliance for white-label deployments

• Better transparency during account registration

Display Wallet Recharge Reason in Invoice

Wallet recharge invoices now include the specific reason for the recharge to improve billing clarity and reconciliation.

Previously, invoices only displayed the amount paid and labeled the transaction simply as “Wallet Recharge”, without explaining the purpose of the recharge.

What’s New

Recharge Reason in Invoice Summary

Invoices now display the reason associated with the wallet recharge, such as:

• “Amount added to pay Chat Agent subscription”

• “Additional amount purchased separately”

Improved Transaction Transparency

While the dashboard already tracked detailed wallet transaction reasons, this information was not previously included in the invoice/receipt. The invoice template has now been updated to display this information directly.

Impact

• Provides clearer financial records for users

• Improves reconciliation between wallet transactions and invoices

• Helps users understand the purpose of each wallet recharge directly from their receipts

collected_at Field Restored in Lead Webhook Payloads

The collected_at field has been restored in lead webhook payloads.

Previously, lead collection webhooks were arriving without the collected_at timestamp, which indicates when the lead was captured. This field had been included in earlier payloads but was recently missing.

What’s Fixed

collected_at Timestamp Restored

Lead webhook payloads once again include a non-null collected_at timestamp, ensuring the time of lead collection is properly communicated in webhook events.

Example payload snippet:

{
  "webhook_type": "lead_collected_via_form",
  "collected_at": "2026-03-05T12:22:27Z"
}

Impact

• Restores the expected webhook payload structure

• Ensures downstream systems can accurately track when a lead was collected

• Improves compatibility with integrations that rely on the collected_at timestamp for processing or analytics

Conversation Location Display Fix

Resolved an issue where conversation locations in the Conversations tab were displaying incorrectly.

Issue

Previously, the location field for conversations showed "Unknown, Unknown" instead of displaying the user's actual city and country.

Fix

The location resolution and display logic has been corrected so that conversations now properly show the user's detected city and country based on their IP information.

Impact

• Restores accurate location visibility in conversations

• Improves conversation context and analytics

• Ensures location data displays correctly instead of defaulting to "Unknown, Unknown"

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

This release delivers major UI enhancements, API improvements, cost optimisations, billing safeguards, and important stability fixes across Conversations, Sub-Accounts, Webhooks, Chat Agents, and Voice Agents.

Chat Widget (ReactJS) & Appearance Revamp

The Chat Widget has been fully migrated to ReactJS, delivering improved performance, flexibility, and easier integration into modern frontend applications. Alongside this upgrade, the Appearance tab has been completely redesigned to provide deeper customisation capabilities.

What’s New:

ReactJS-Based Chat Widget
Rebuilt the widget using React for better maintainability, scalability, and seamless integration into React-based platforms.

Two Customisation Modes
Introduced dual styling options to support both non-technical and advanced users:

• UI Settings Mode – Design your widget visually using structured customisation panels.

• Custom CSS Mode (New) – Apply fully custom CSS for advanced styling control.

Expanded UI Controls
The redesigned Appearance tab now includes enhanced configuration sections such as:

• Content & Messages

• Layout & Display

• Features

• Automation & Timing

• Typography

• Colors & Styling

• Images & Assets

Live Preview Panel
Real-time preview updates allow users to instantly visualise changes before saving.

Impact:

Agencies and businesses can now create highly customised, brand-aligned chat widgets, whether through simple visual controls or advanced CSS styling without limitations.

Sub-Accounts Dashboard Redesign

The Sub-Accounts dashboard has been completely redesigned with improved structure, enhanced visibility, and new organisational tools for agencies managing multiple clients.

What’s New

Redesigned Card & List View

• Cleaner visual hierarchy

• Improved spacing and alignment

• Faster account scanning and management

Resource Summary Overview

A new summary section at the top displays total allocated resources across the agency:

• Chat Agents

• Voice Agents

• Character Storage

• Wallet Balance

This provides an instant snapshot of overall allocation and usage.

Advanced Filtering System

A structured filter panel now allows segmentation by:

• Status

• Plan

• Billing Cycle

Includes quick Apply and Reset actions for efficient account management.

Folder Organisation System (New)

• Create custom folders with colour tags

• Group related clients

• Toggle between “All” and specific folders

• Combined search + folder navigation

Designed to support agencies managing large client portfolios.

Improved Account-Level Visibility

Each sub-account card now clearly displays:

• Subscription status

• Billing type (Direct Billing / Agency Billing)

• Resource allocation and usage

• Account balance

Reduces the need to open individual accounts for basic checks.

Enhanced Add / Edit Sub-Account Drawer

The side drawer has been restructured for clarity and efficiency:

• Clear separation of:

  • Account Details

  • Resource Allocation

  • Billing & Trial

  • Permissions

• Improved field grouping

• Cleaner toggle controls

• Standardised spacing and typography

Impact:

Sub-account management is now faster, clearer, and more scalable for agencies. Resource allocation, billing structure, and account health can be reviewed at a glance.

Human / Agent Toggle API Exposure

The Human / Agent toggle available within the Conversations UI has now been fully exposed as a public API endpoint.

Previously, switching a conversation between Chat Agent and Human modes could only be done manually through the dashboard interface. This functionality is now accessible programmatically via API.

What’s New:

• Added a new API endpoint to update conversation mode

• Supports switching a conversation to:

  • human – hands off the conversation to a human operator

  • agent – returns control back to the AI Chat Agent

• Requires conversation_id and mode as parameters

• Returns the updated conversation state

• Mirrors the exact behaviour of the UI toggle

Conversation Title Generation Optimisation

Conversation titles are now generated within the same OpenAI request instead of being created through a separate API call.

Previously, title generation required an additional OpenAI request after the assistant response. This has now been optimised by embedding title generation directly into the primary conversation prompt, returning both the assistant response and the conversation title in a structured output.

What’s New:

• Title generation is handled within a single OpenAI request

• Eliminates the need for a separate title-generation API call

• Introduced a new “LLM-Driven Conversation Titles” toggle in the Chat Agent Settings tab

• When enabled, titles are automatically generated for conversations

• When disabled, conversation titles will not be generated via the LLM

This gives administrators full control over whether title generation is active.

Impact:

• Reduces overall API usage and token consumption

• Improves response efficiency by removing redundant requests

• Simplifies the conversation workflow

• Provides optional control over automatic title generation through settings

This update improves cost efficiency while maintaining flexibility and control at the agent level.

Custom Function History JSON Formatting

Improved formatting of JSON response payloads in the Custom Function history table.

Impact:

• Enhanced readability

• Improved debugging experience

• Cleaner structured output display

Webhook History Pagination Fix

Resolved an issue where paginated webhook history API responses caused incomplete UI records.

What’s New

• Fixed pagination handling

• Added a new “Resend” action column in the webhook history table

Impact:

• Complete webhook visibility

• Improved event recovery workflows

• Greater reliability in webhook management

Lead Collection Form Enhancement

Enhanced the Lead Collection Form with additional structure and visibility.

What’s New

• Added support for Dropdown fields with configurable options

• Exposed Chat Widget form submissions directly to agents

Impact:

• More structured lead capture

• Improved context during conversations

• Better sales and support workflows

SaaS Package Configuration Fix

Resolved missing Knowledge Base storage allocation for Voice Agents in SaaS package configuration.

Impact:

• Correct resource allocation

• Prevented feature limitations for Voice Agents

• Improved package reliability

Accessibility Improvement

Added a title attribute to the embedded chat iframe for improved accessibility compliance.

Impact:

• Better compatibility with assistive technologies

• Improved accessibility standards

Voice Agent Billing Improvement

Added a “Clear Dues” button for Voice Agents with pending phone number payments.

Impact:

• Faster resolution of outstanding payments

• Improved billing transparency

• Simplified account management

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

What’s New:

• Added Calendly appointment scheduling support to AI Voice Agents.

• Voice agents can now book meetings directly during inbound or outbound calls.

Impact:
Enables hands-free scheduling via voice, improving automation and lead conversion for call-based workflows.

Typing Indicator Experience for Meta Channels

What’s New:

• Improved typing indicator behavior for WhatsApp and Instagram conversations.

• Prevents unnecessary or misleading typing signals.

Impact:
Creates a more natural and polished chat experience on Meta platforms.

MasterChat Styling Restored in White Label Settings

What’s New:

• Restored MasterChat customization options under Agency → White Label → Custom Styling.

• Agencies can again configure MasterChat name and branding assets.

Impact:
Restores branding control for white-labeled experiences and agency customization.

Custom Parameters: JSON Saving & Array Handling Improvements

What’s New:

• Fixed an issue where custom parameter names were altered after saving JSON.

• Added native support for array-type parameters.

• Improved sync between JSON View and Form View, preventing empty or mismatched fields.

• Automatically assigns a default token name for non-primitive parameter types.

Impact:
Makes custom tool and function configuration more reliable, predictable, and developer-friendly.

Restored Tooltip Messages for Agency Settings

What’s New:

• Restored missing tooltip help text across key agency configuration areas.

Impact:
Improves clarity and reduces confusion when managing advanced settings.

Model Upgrade: Grok Migration

What’s New:

• Migrated from deprecated grok-2-1212 to grok-3.

Impact:
Ensures continued model support, improved performance, and future compatibility.

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

Improved clarity in API key setup by showing accurate helper messaging based on the key’s state.

What’s New:

• Displays a meaningful helper message when no API key is entered (e.g., prompting users to add a key).

• Encryption-related messaging is now shown only after a valid API key is successfully saved.

• Applied consistently across OpenAI, XAI, Anthropic, and ElevenLabs key tabs (where applicable).

Impact:
Eliminates misleading messaging during initial setup, reduces user confusion, and guides users more effectively when configuring API keys.

Save Button UI Improvements

Updated the save and update actions on the dashboard to be more visible and intuitive.

What’s New:

• Save and update buttons are now larger and more prominent.

• Buttons have been repositioned under the main configuration section instead of the preview card.

• Improved consistency with other dashboard actions.

Impact:
Enhances clarity and usability, making it easier for users to confidently save changes without missing critical actions.

Increased Chat Agent Add-on Price Limit

Updated pricing validation for Chat Agent add-ons in subscription packages.

What’s New:

• Increased the maximum allowable price for Chat Agent add-ons beyond the previous limit.

• Updated validation rules to support higher price values without errors.

Impact:
Provides agencies with greater pricing flexibility, enabling more customizable monetization strategies and premium service offerings.

Sub-Account Wallet Balance Management Improvements

Enhanced wallet management controls for agency users.

What’s New:

• Previously, agency users could only increase sub-account wallet balances.

• Agency users can now both increase and decrease sub-account wallet balances.

• Added proper handling for balance adjustments, corrections, and reversals.

Impact:
Restores full financial control for agencies, eliminates a key limitation in balance management, and helps prevent billing errors and inconsistencies.

Automated Signup Email Improvements

Resolved issues affecting automated signup email templates.

What’s New:

• Fixed a broken custom-domain link in the opening line of signup emails.

• Improved handling of the name token so greetings no longer appear empty when a name is unavailable.

Impact:
Delivers a smoother, more professional onboarding experience and prevents confusion caused by broken links or incomplete greetings.

Notification History Tracking for Webhooks & Emails

Added visibility into webhook and email notification activity.

What’s New:

• Introduced history tracking for email notifications and webhook URLs under
  Analytics → AI Agent → Summary Settings.

• History behavior is consistent with existing tracking in the Leads section.

Impact:
Makes it easier to debug, audit, and validate notification workflows by providing clear historical context for webhook and email activity.

Free Trial Settings Sync Improvements

Resolved an issue where free trial configuration changes were not reflecting in sub-accounts.

What’s New:

• Free Trial Settings updates made at the agency level now correctly propagate to all associated sub-accounts.

• Sync behavior is consistent for both existing and newly created sub-accounts.

Impact:
Ensures predictable and consistent trial behavior across accounts, reducing confusion and preventing misaligned trial experiences for end users.

Wallet Transaction Tracking for Messaging

Improved transaction visibility for message-based usage.

What’s New:

• Wallet transactions are now recorded for messages sent by agency-owned bots.

• Includes messaging usage when agencies use their own API keys.

Impact:
Delivers consistent and complete wallet transaction history, improving billing transparency, reporting accuracy, and usage auditing across agency and sub-account messaging.

SaaS Configurator: “Most Popular” Package Control

Added manual control over which subscription package is highlighted as Most Popular.

What’s New:

• Agencies can now manually select any package as the “Most Popular” in the SaaS Configurator.

• If no package is selected, the system continues to default to the middle package.

• The selected setting is saved and consistently reflected in the customer-facing package view.

Impact:
Gives agencies greater flexibility in presenting pricing plans, helping highlight preferred offerings and improve conversion rates.

Dashboard Token Usage Visibility

Expanded usage analytics across agents and sub-accounts.

What’s New:

• Token usage is now displayed per AI agent and per sub-account directly on the dashboard.

• Usage data integrates with existing dashboard filters for easier drill-down and analysis.

Impact:
Gives agencies clearer insight into how token usage and costs are distributed across agents and sub-accounts, supporting better monitoring and cost control.

Conversation & Timestamp Consistency Improvements

Improved consistency and clarity across AI conversation views and APIs.

What’s New:

• Standardized timestamp handling across platforms to ensure consistent dates and times.

• AI conversations now display the actual lead’s first and last name instead of generic visitor labels.

• Fixed the missing conversation title (chat summary) field in the Conversations API.

Impact:
Provides clearer conversation context, consistent reporting across platforms, and more reliable data for integrations and downstream systems.

Meta Story Reply Handling Improvements

Improved bot behavior for Meta (Instagram/Facebook) story interactions.

What’s New:

• Bots no longer respond to story replies that contain emojis only.

• Prevented automated replies to story reactions that are not true conversation starters.

Impact:
Reduces unnecessary bot responses, prevents alert fatigue for clients, and ensures bots engage only in meaningful conversations, improving user experience on Meta platforms.

Live Support Option

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

Stammer now includes fully interactive AI Voice Agents capable of answering and making calls with natural, human-like responses. These agents can qualify leads, book appointments, answer FAQs, and operate 24/7 with no delays or missed calls.

Live 24/7 Call Handling

Voice agents instantly pick up inbound calls at any hour. Ideal for teams looking to eliminate voicemail, missed leads, and after-hours gaps.

Natural, Human-Sounding Voices

Choose from multiple lifelike voice options. Each voice is designed to sound smooth, conversational, and context-aware; not robotic.

Simple 60-Second Setup

Launch a voice agent with minimal configuration:

• Connect your number or Twilio

• Name your agent

• Add instructions

• Provide training data
  You’re fully live in under a minute.

Enhanced Knowledge Base Training for Voice

Voice agents instantly learn from your knowledge base:

• File uploads

• Q/A sections

• Text snippets

• Website content
  They use this data to give accurate, context-aware answers during calls.

Full Call Control & Lead Automation

Agents can:

• Answer inbound calls

• Initiate outbound calls (coming soon)

• Transcribe conversations

• Generate call summaries

• Capture lead details
  All data can be emailed or sent to your preferred systems.

Sell Voice Agents to Clients

Agencies can now offer voice agent services as billable add-ons. Create packages, set margins, include usage pricing, and resell voice capabilities with no technical overhead.

Inbound Calling Upgrades

Agents handle:

• New lead inquiries

• After-hours calls

• FAQ responses

• Appointment scheduling

• Lead qualification

• Summary emails
  Each call is logged with a transcript, summary, and lead info.

Outbound Calling (Coming Soon)

Upcoming features include:

• Automated lead follow-up

• Appointment confirmation

• Customer reactivation

• Outbound sales workflows
  With results sent directly to your email or CRM.

Live Examples Added

To help you understand real-world use cases, Stammer now includes example voice agents for:

• Dental

• Legal

• Plumbing

• Real estate

• Healthcare

• Fitness
  And additional industries.

Why This Matters

Voice dramatically increases answered calls, captured leads, and booked appointments for both you and your clients. Stammer 2.0 gives agencies a new revenue stream and a powerful automation channel with minimal setup.

Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

The filters in the SaaS configurators: Active, Legacy, and Private — now include short, informative descriptions to clarify their purpose and impact on visibility and usage.
This enhancement improves clarity and overall user experience by helping users understand what each filter represents.

2. Fixed Chatbot Allocation Mismatch

Resolved an issue where the allocated chatbot count did not match the actual sub-account allocations, ensuring accurate display and usage tracking across accounts.

3. API Docs: Sample Payloads Updated

Resolved an issue where sample payloads in the API Docs were displaying “None” instead of actual example values.
The payload examples now show realistic request and response data, making the documentation clearer and easier to understand for users.

4. Increased OTP Expiration Time for Email Updates

Extended the OTP validity period to 180 seconds when updating an account email, giving users more time to complete the verification process without re-requesting codes.

5. Fixed Masterchat UI Glitch on AI Replies

Addressed a UI glitch in Masterchat that occurred when an AI agent sent a message, ensuring smooth transitions and consistent layout updates during conversations.

6. Investigated & Fixed Post-Trial Payment Status Issue

Fixed an issue where a sub-account remained marked as Paid even after a failed post-trial payment. Payment state logic now updates correctly in all cases.

7. Enhanced wait_for_browser Stability

Improved the wait_for_browser function within the Advanced Scraping module to provide more reliable page readiness checks and reduce errors caused by partially loaded DOMs.

What’s New:

• Added support for additional wait conditions, including DOM element selectors, network idle states, and JavaScript variable checks.

• Introduced timeout and retry logic for dynamic content handling.

• Enabled the option to inject custom wait logic for advanced use cases.

8. SaaS Configurator – Improved Filter & Package Tier Handling

Enhanced the SaaS Configurator experience by addressing filter persistence and improving package visibility.

What’s New:

• The selected filter (Active, Private, Legacy, etc.) now remains active when navigating between package details and the main list, ensuring smoother navigation.

• Package tier labels (e.g., Free, Starter, Pro) are now clearly displayed on each package card for quicker identification.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

What’s New:

• Usage-Based Message Billing:
  Instead of a flat rate for every AI message, pricing now adjusts based on the size and data usage of each message, just like OpenAI’s per-token model.

• Automatic % Margin Conversion:
  All existing agency pricing has been automatically converted to an equivalent percentage margin, ensuring your customers continue paying roughly the same as before.

• Consistent Profit Margins:
  You now earn a steady profit percentage on every message, whether it’s short or long. For example, if your cost is $0.01 and your margin is 100%, your customer pays $0.02 automatically adjusted for message size.

• Fair Pricing for Everyone:
  Around 95% of messages will still cost roughly the same. Short messages will cost a bit less, while longer, more data-heavy messages will cost slightly more.

• Custom API Key Support:
  Agencies using their own OpenAI API keys will continue being billed directly by OpenAI. The same percentage margin model applies seamlessly.

Why This Matters:

Previously, every message no matter how short or long was billed at the same rate. Now, pricing reflects actual message size, ensuring fairness and better alignment with real usage while keeping profit margins consistent for agencies.

Impact:

• No manual changes needed, your account has already been updated automatically.

• Existing plans and packages remain exactly the same.

• Customers won’t see any billing disruption or major changes unless they send exceptionally large or data-heavy messages.

What You Can Do:

• Review your % margin in the dashboard, it’s already been configured to match your previous pricing.

• Adjust your margin anytime to control your profit percentage.

• Continue using your existing setup, nothing else changes in functionality or workflow.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

Significant backend upgrades have been made to boost performance, scalability, and stability.

What’s New:

• Optimized nearly 65% of system queries, reducing redundant database operations and overall DB load.

• Improved execution time for frequently accessed endpoints.

• Enhanced system responsiveness and reliability, especially for high-traffic environments.

Enable Package Downgrade for Sub-Accounts

Sub-account users can now downgrade their subscription packages directly from the dashboard, no need to contact customer support.

What’s New:

• Added functionality allowing sub-account users to downgrade packages via the platform UI.

• Removed the previous restriction and “Downgrading not allowed” error message.

• Implemented proper validation and confirmation flow to ensure smooth and secure downgrades.

Added "Delete Conversation" Option & Display AI Agent Name in Masterchat

A new Delete Conversation option has been added to Masterchat, along with agent identification by showing agent name at the top of each conversation.
What’s New:

• Users can now delete specific conversations directly from the Masterchat interface.

• Each conversation now displays the AI Agent name for better context and management.

• Enhances usability for agencies managing multiple bots.

Simplified Vault Key Settings

Simplified the API Key setup process in the Vault tab (Agency Settings).
What’s New:

• “Verify” and “Save” actions have been merged into a single button.

• Streamlines workflow and minimizes confusion when adding or updating API keys.

Debug Scheduler Logs for Auto-Scraping & Training

Enhanced scheduler reliability for auto-scraping and training jobs.
Fix Highlights:

• Resolved issues where threads were not starting or completing correctly.

• Fixed scenarios where chatbot training was not triggered post-scraping.

• Improved visibility in scheduler logs for debugging and monitoring.

View Input/Output Token Usage per Message

All users can now view token usage for each message.

What’s New:

• Added per-message input/output token count in the Dashboard Widget.

• Provides better transparency into how tokens are being consumed during conversations.

• Helps users monitor and optimize model performance and usage costs.

Include Conversation ID in Custom Function Payloads

Each custom function call now includes a unique Conversation ID (UUID) in the payload.
What’s New:

• Allows more precise tracking of context-specific function executions.

• Simplifies debugging and correlation between chatbot messages and API events.

Enhanced Package Creation Flow (Private & Sub-Account Visibility)

Improved error handling and clarity when creating or modifying packages of the same tier.
What’s New:

• Clearer error message displayed when trying to enable “Show package to sub-accounts” for an existing package tier.

• Added direct link option for private packages, allowing easier distribution and controlled access.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

We’ve completely redesigned the SaaS Configurator to deliver a cleaner, more intuitive, and efficient setup experience.

Here’s a complete walk-through of what’s been updated:

Added AED Currency Option to Platform

We’ve added support for the AED (United Arab Emirates Dirham) as a selectable currency throughout the platform.
This enables more accurate pricing, billing, and localisation for users operating in the UAE region.

Retrieve Sub-Account Chat Agents via API

You can now retrieve all Chat Agents belonging to a specific Sub-Account using a newly introduced API endpoint.

What’s New:

• Added a GET API endpoint for fetching chat agents from any sub-account under an agency.

• Simplifies integration workflows for developers who need to list or manage agents programmatically.

• Fully documented and available in the API reference.

API Endpoint:

GET https://app.stammer.ai/en/api/v1/subaccounts/chat-agents/your_subaccount_id/

Use Case Example:
Fetch all chat agents for a given sub-account to display, manage, or automate configuration updates in your agency tools.

List Sub-Accounts via API

A new API endpoint has been introduced to allow agencies to programmatically retrieve all Sub-Accounts along with their basic details and IDs.

What’s New:

• Added a GET API endpoint for listing all Sub-Accounts within an Agency.

• Returns key information such as Sub-Account ID, name, and associated metadata.

• Enables easier automation, reporting, and integrations for agency-level management.

API Endpoint:

GET https://app.stammer.ai/en/api/v1/subaccounts

Use Case Example:
Integrate this endpoint to automatically fetch and display Sub-Account data in your internal dashboards or synchronize with external tools.

Fix – Agency Owners and Permitted Users Unable to Create or Update Sub-Accounts

Resolved an issue where Agency Owners and permitted users were temporarily unable to create or update sub-accounts.

Fix Highlights:

• Restored proper permissions for both agency-level and delegated users.

• Ensures seamless sub-account creation and editing functionality.

Fix – Embedded Agents in Private Chatbots Showing "Access Denied from This Domain" (403 Errors)

Fixed a bug where embedded agents configured as Private Chatbots were incorrectly displaying “Access Denied from This Domain” or 403 errors, even for authorised domains.

Fix Highlights:

• Updated domain validation logic for embedded private chatbots.

• Ensures proper functionality for whitelisted domains only.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

We’ve reorganised agency-level settings to make them easier to find and manage.

What’s New:

• All major agency settings are now consolidated under a single Agency tab in the profile section.

• Includes: Dashboard Settings, Custom Domain, Custom Styling, Stripe Integration, and Account Notifications.

• Provides a more intuitive and accessible layout for managing agency-wide configurations.

Sub-account Filtering by Billing Category

We’ve made it easier to manage sub-accounts by introducing filtering and sorting options based on billing status.

What’s New:

• Added a filtering dropdown in the Sub-accounts view to group or filter sub-accounts by billing category:

  • Paid

  • Expired

  • Directly Billed

• Optionally, each sub-account card can now display its billing status for quicker visibility.

Why it matters:

• Simplifies navigation and management for agencies with many sub-accounts.

• Provides clear categorisation of billing statuses at a glance.

• Improves efficiency and reduces admin overhead.

"Save Settings" in Knowledge Base > Websites Tab

Previously, changes made in the Websites tab of the Knowledge Base could only be saved by starting a new scraping process.

What’s New:

• Added a dedicated “Save Settings” button to allow saving updates without initiating scraping.

• Improves usability by letting users adjust and persist configurations independently.

• Provides more control and flexibility in managing website-based knowledge sources.

Improved Currency Change Workflow for Agencies

We’ve simplified the process for agencies to change their currency settings. Previously, agencies had to delete all SaaS packages, remove paid subaccounts, and ensure their wallet balance was zero before being allowed to switch currencies. This process was confusing and restrictive.

What’s New:

• Wallet balance handling: Agencies can now change currency even with a wallet balance.

  • A confirmation warning is shown:
    "Changing your currency will convert your existing wallet balance to the new currency. Please confirm before proceeding."

  • Wallet balances are automatically converted on the backend.

• Clearer instructions: When the currency field is disabled, the system now shows clearer guidance:

  "To change your currency, please delete all paid subaccounts and SaaS packages. Wallet balance will also need to be cleared or will be auto-converted if you proceed."

Font Consistency in React Pages

Resolved font inconsistencies across multiple React pages.

Fix Highlights:

• Unified typography across app components.

• Improves professional appearance and readability.

Prompt Template Dropdown – Increased Width

Improved readability for long prompt template names.

What’s New:

• Dropdown width expanded to fit longer template titles.

• Prevents text clipping and overlapping issues.

"Select All" in Scraped URLs Tab

Fixed bug where the “Select All” checkbox did not function on initial load.

Fix Highlights:

• Checkbox now correctly selects all scraped URLs.

• Improves bulk management of scraped data.

Clarifying VAT Text in SaaS Configurator (EU Only)

We’ve improved clarity around VAT settings in the SaaS configurator to help reduce confusion.

What’s New:

• Added an explanatory note to the VAT configuration section:

  “These VAT settings apply only to customers based in the EU.”

Why it matters:

• Helps non-EU agencies understand that VAT configuration is not required for them.

• Reduces support requests and prevents unnecessary misconfigurations.

• Provides a clearer, more intuitive experience when setting up SaaS billing.

Claude Model Performance Investigation

We’ve addressed performance issues reported with Claude models (including Claude 3.5 Haiku and Claude 3.7 Sonnet).

What was happening:

• Conversations were occasionally getting “stuck.”

• Responses were slower than expected, leading to a degraded user experience.

What’s fixed:

• Improved response handling to reduce delays.

• Optimized processing to prevent conversations from freezing.

Remove Custom Domain from Agency Settings

Agencies can now remove previously added custom domains from their dashboard.

What’s New:

• Greater flexibility in domain management.

Clickable URLs in Knowledge Base Search

Knowledge base search results now display clickable URLs.

What’s New:

• Faster navigation directly from search results.

Wallet/Storage CSV Report Improvements

We’ve improved the way CSV reports are delivered from the Wallet Transactions and Storage Transactions tables (found under Billing Usage & Settings → Usage tab).

What’s Changed:

• CSV exports are now sent directly to the user who requested them.

• The agency owner’s email is always CC’d on every CSV export, regardless of who initiates the request.

Chat Widget UI Fixes

We’ve addressed two minor UI issues in the Chat Widget and Knowledge Base for a smoother user experience.

Fixes:

1. Broken Avatar Image in Appearance Tab

   • Before: If no avatar image was uploaded, a broken image icon appeared in the Chat Widget → Appearance → Chat Icon section.

   • Now: A proper placeholder avatar is displayed (or the image element is hidden) when no image is uploaded.

2. Unformatted Character Count in Scraped URLs

   • Before: In Knowledge Base → Scraped URLs, the total character count appeared unformatted (e.g., 1234567).

   • Now: Character counts are comma-separated for better readability (e.g., 1,234,567).

Mobile UI Glitch – Usage Details Dropdown

We’ve fixed a layout issue affecting the “View Usage Details” button in mobile view.

What was happening:

• On smaller screen devices, the button was misaligned and overlapped with surrounding UI elements.

• This created both a visual inconsistency and a usability problem.

What’s fixed:

• The “View Usage Details” button is now properly aligned and responsive across all mobile screen sizes.

• Improved spacing ensures no overlap with other UI elements.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

The Leads tab within the AI Agent section now includes detailed tracking for email notifications related to lead capture.

What’s New:

• Added Test Notification option to validate email alerts for lead collection.

• Introduced Email Notification History to view when and what notifications were triggered.

• Enhances visibility for both testing and live lead capture workflows, making troubleshooting easier.

  

Template Prompts in AI Agent Settings

Users can now quickly configure their AI agents using prebuilt prompt templates pulled directly from the marketplace.

What’s New:

• A new dropdown above the prompt field in the Settings tab lets users select from available prompt templates.

• Agencies: Access prompt templates listed in Stammer’s public marketplace.

• Sub-accounts: View and use templates shared via their Agency’s private marketplace.

• Streamlines agent setup, encourages prompt reuse, and ensures consistency across teams.

Enable Deletion of Failed or Scraped URLs

Users can now delete failed or scraped URLs in the Knowledge Base.

Fix Highlights:

• The "Delete" button is no longer disabled.

• Empowers better control over knowledge base management.

Move AI Agent into Folder

Users can now organize their AI agents more easily by moving them into folders via the dashboard.

API Endpoints for Widget Configuration and Custom Functions

Developers can now manage chatbot customization and custom function logic directly via API endpoints and updated UI sections.

Custom Functions API

Now accessible via both the UI and API, allowing users to:

• List Custom Functions
  Retrieve a list of all available custom functions for a chatbot.

• Create Custom Function
  Add new custom logic via the API.

• Update Custom Function
  Modify existing functions easily.

• Delete Custom Function
  Clean up unused or deprecated logic.

This addition improves workflow automation and simplifies backend integration for developers.

Chat Widget Configuration API

• View Chat Widget Configuration: GET /api/v1/chatbots/<bot_id>/chatwidget-configuration/

• Update Chat Widget Configuration: PUT /api/v1/chatbots/<bot_id>/chatwidget-configuration/

Customize and control chatbot widget visuals programmatically using structured payloads (e.g., icons, avatars, font, background, gradient, etc.).

"View Tutorial" Button for Each Custom Function Template

A "View Tutorial" button is now available for each Custom Function Template to help users quickly understand and implement them effectively.

What’s New:

• 📘 Button Label: View Tutorial

• 🔗 Opens a tutorial link in a new tab for step-by-step guidance on how to use the selected custom function template.

Toggle for Auto-Refresh in Embedded Chatbots

Users can now disable automatic conversation refresh in embedded chatbot widgets to maintain chat continuity.

What’s New:

• 🛠️ Optional setting added to prevent the embedded bot from auto-refreshing during inactivity.

• 💬 Preserves ongoing conversations and prevents loss of context due to unwanted page or widget refreshes.

• ⚙️ New setting location: Found under the Integration tab, where the embedded code is displayed.

• 🚫 Disabled by default to protect chat sessions unless explicitly enabled.

  

Single-Click Menu Update

Improved navigation by changing the multi-click dropdown menus to work with a single click.

Private Embed Privacy Enforcement Fix

Resolved a bug where private embed chatbots were still accessible outside of the specified domain.

Fix Highlights:

• Enforces strict access control as defined in embed settings.

Updated Portuguese Translations

Updated the Portuguese translation files across the app based on the latest content provided by the client.

Deleted KB URLs Still Appearing in Search

Resolved a bug where deleted Knowledge Base URLs were still appearing in the KB search tab.

Human Handoff Webhook Triggered Twice

Resolved an issue where the webhook for human handoff was firing twice instead of once.

Fix Highlights:

• Ensures only a single webhook trigger per handoff event.

• Reduces noise and duplicate data in webhook consumers.

Duplicate & Missing Messages in Masterchat

Fix applied to address:

• Duplicate messages appearing in conversations.

• Certain conversations not being displayed in Masterchat view.

Duplicate lead_data Triggers

Investigated and fixed the issue causing the lead_data trigger to execute twice, which was leading to duplicate email notifications.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

(NEW) ☎️ Daily Support Calls
Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
Link to join: https://stammer.ai/daily-support-call

Conversation Read/Unread Indication for Each User
We've introduced individual read/unread markers to human handed off conversations, helping teams stay aligned and responsive.

What’s New:

• Each user now sees their own read/unread status per conversation for better collaboration.

• When human handoff is triggered and the end-user sends new messages, a blue dot will now appear, indicating there are unread messages in the conversation.

Enhances visibility into which conversations need attention and ensures no message goes unnoticed during live support or escalations.

Ability to Pin AI Agents to Top of the List
You can now pin key AI agents to the top of your agent list for quicker access.

What’s New:

• Pinned agents are user-specific and remain pinned across sessions.

• Enables faster navigation by keeping high-priority agents always visible.

• Simplifies management without needing to rename or manually reorder agents.

This enhancement helps streamline workflows for users managing multiple agents.

Added API Endpoint for Adding Text into Knowledge Base
You can now add text data directly to the Knowledge Base via API.

What’s New:

• New API endpoint for programmatic additions of KB entries.

• Supports automation and faster knowledge ingestion.

  

Clone Agents from Sub-Account to Agency
You can now clone AI agents from a sub-account directly to the agency.

What’s New:

• Preserves configuration, logic, and training data.

• Enables faster reusability and standardisation of agents across teams.

  

Added PLN as a Supported Currency Option
We’ve added support for the Polish złoty (PLN) in billing settings.

What’s New:

• Agencies can now price packages in PLN.

• Subscribers will be billed using the selected currency.

✅ Fixes & Enhancements

Sorted Webhook History Table by Latest First & Add Export Option
We’ve enhanced the usability of the Webhook History table to make data access faster and more actionable.

What’s New:

• Webhook entries are now sorted by most recent first, making it easier to view the latest events.

• Added an Export option to the Webhook history table, allowing users to download records for analysis or reporting.

• Improves visibility and flexibility for debugging, auditing, or campaign follow-ups.

Unable to Launch AI Agent to Marketplace – No Agents Displayed
Resolved an issue where the AI Agent dropdown in the launch-to-marketplace flow was failing to populate.

Fix Highlights:

• Fixed the broken dropdown that showed no agents available for selection.

• Resolved related search and filtering issues preventing agent visibility.

• Agents can now be correctly listed and selected when publishing to the marketplace.

Avoid Creating $0 Transactions on Sub-Account Updates
We’ve eliminated unnecessary $0 transaction records when sub-account billing details are updated.

Fix Highlights:

• Reduces clutter in the transaction log.

• Only meaningful financial transactions are recorded going forward.

Show All Conversations After Clearing Search in Conversation Tab
Resolved an issue where conversations were not reloading after clearing a search filter.

Fix Highlights:

• Clearing the search input now restores the full conversation list immediately.

WhatsApp Deleted Conversations Not Showing in Filter/API
Fixed a bug that excluded deleted WhatsApp conversations from both the Deleted Conversations filter and API.

Fix Highlights:

• Deleted WhatsApp threads now appear correctly in both frontend and API endpoints.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

Notify Users on Grok & Claude API Failures
To improve transparency and troubleshooting, we’ve added real-time failure notifications for AI model errors.

What’s New:

• Users are now notified when Grok or Claude model API calls fail.

• Error messaging provides context and suggested actions to recover.

• Helps users quickly understand and resolve integration issues.

Show Answer Sources in Conversations
We now display the source of responses in the conversation view for better clarity and trust.

What’s New:

• Each AI-generated answer now includes its source (e.g., Knowledge Base, Custom Functions).

• Makes it easier for users to trace back and verify information.

• Particularly helpful for training, debugging, and auditing use cases.

Display Request Type in AI Agent Custom Functions Table
The custom functions table for AI agents now shows the HTTP request type for each function.

What’s New:

• Displays GET, POST, PUT, etc., next to each function entry.

• Helps developers and admins quickly understand function behavior at a glance.

✅ Fixes & Enhancements

Fix Long URLs Breaking UI in Scraped URLs Tab
Improved the layout in the Scraped URLs tab to better handle long links.

Fix Highlights:

• Added word-breaking and text wrapping.

• Prevents overflow and layout distortion.

Incorrect Dates Displayed in Conversations Sidebar
Resolved a bug that caused misleading or incorrect timestamps to appear in the conversation sidebar.

Fix Highlights:

• Corrected date formatting and logic for display.

• Reduces user confusion and improves conversation tracking.

Improve Error Handling for Existing Users Trying to Sign Up Again
Improved backend handling and messaging when a user attempts to sign up with an already existing account.

Fix Highlights:

• Clear error message now shown: “This email is already registered.”

• Prevents partial signups and redundant account creation attempts.

Fixed UI Labels in Usage Details Dropdown (Wallet Usage)
Standardised the language and alignment of labels in the usage breakdown dropdown.

Fix Highlights:

• Corrected inconsistent or unclear label names.

• Improved layout for readability and clarity.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

Updated Support Tab for Agency Users
The support tab has been redesigned specifically for agency users.

What’s New:

• Clearer layout with relevant help links and resources.

• Tailored to agency workflows and support needs.Redirected Old Conversation Path to the New Conversation Path
  Old URLs pointing to outdated conversation paths are now automatically redirected to the new route.

What’s New:

• Redirect logic ensures seamless navigation and backward compatibility.

• Users accessing older links will land directly on the correct updated conversation view.

  

Option to Export Storage Transactions
You can now export all storage transaction data for record-keeping or reporting purposes.

What’s New:

• Export functionality similar to wallet transactions.

• Supports CSV format for easy integration with external tools.

  

Automated Conversation Summary Based on Widget Inactivity
We’ve introduced an automated system to summarise conversations when user inactivity is detected in the chat widget, helping streamline context recovery for returning users and support agents.

What’s New:

• The system now monitors inactivity and triggers a summary generation to keep conversations fresh and focused.

• A conversation refresh is automatically triggered under the following conditions:

  • User closes the browser tab.

  • Continuous inactivity for a defined duration.

  • User switches to another browser tab and remains inactive for 15 minutes.

  • Tab remains open but receives no interaction for 30 minutes.

  • User navigates to a different route (route change detected).

This ensures that long-idle conversations don’t create confusion or lag and helps agents quickly understand context when picking up older chats.

Language Translations in React
We’ve extended multilingual support by translating all newly updated tabs and sections in the dashboard.

What’s New:

• All recently updated dashboard tabs are now fully translated into the supported languages.

• Maintains consistency across the interface for users working in non-English environments.

• Enhances the overall usability and accessibility of the platform for global audiences.

This ensures that both existing and updated UI elements are aligned with your selected language preference.

Redirected Old Conversation Path to the New Conversation Path
Old URLs pointing to outdated conversation paths are now automatically redirected to the new route.
What’s New:

• Redirect logic ensures seamless navigation and backward compatibility.

• Users accessing older links will land directly on the correct updated conversation view.

✅ Fixes & UI Enhancements

Fixed Background Colour of Human Responses in Conversations Tab During Human Handoff
Resolved inconsistent background colour display for messages coming from human agents in handoff mode.

Fix Highlights:

• Background colour now aligns with agent role.

• Enhances conversation readability and role clarity.

Fixed UI of Cards in SaaS Configurator and Subscription Pages
Improved the visual layout of pricing and feature cards.

Fix Highlights:

• Resolved spacing and alignment issues.

• Enhanced responsiveness across device sizes.

Made "Number of Characters Allocated" Read-Only for Users Billed to Agency
We’ve updated the character allocation field for users directly billed through an agency to reflect actual usage and prevent confusion.

What’s New:

• The “Number of Characters Allocated” field is now read-only for agency-billed users.

• Since these users draw from the agency’s shared storage, individual allocation does not apply.

• The field now only displays the number of characters used, not an editable limit.

This change simplifies the UI, avoids misunderstandings, and ensures character usage accurately reflects the agency’s shared limits.

Fixed Cancel Button in Websites Tab of Knowledge Base (AI Agent)
Resolved an issue where the cancel button was not functioning in the Websites section.

Fix Highlights:

• Full button functionality restored.

• Users can now cancel website scraping setups without refresh.

Updated Info Icon Colour to Match Theme
Info icons across the app have been recoloured to better align with the theme and improve accessibility.

Fix Highlights:

• Consistent icon styling across light and dark modes.

• Improves overall UI polish.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

Added VAT to Stripe Links

We’ve added built-in support for VAT/TAX handling in subscription packages created via the Stammer dashboard.

What’s New:

• Agency users can now enable VAT/TAX settings while creating subscription packages.

• Set a custom VAT percentage to apply on top of the base package price.

• The configured VAT is automatically added to all user payments for that package.

• Ensures smoother compliance with tax regulations and transparent billing for subscribers.

This feature gives agencies more control over taxation and improves the clarity of pricing for end users.

Separated Widget Popup Configuration for Mobile and Desktop

Previously, widget popup configuration (e.g., position, size, and behavior) was applied universally across all devices — leading to layout or usability issues when optimizing for one platform but not the other.

What’s New:

• Independent Configuration: You can now define separate widget settings for desktop and mobile views.

• Improved Usability: Avoid layout clashes and ensure optimal chat experiences tailored to screen size and interaction style.

• Customizable Options per Platform: Includes position (left/right), initial popup behavior, size, and more.

  

Added Configurable Toggle for Initial Pop-Up Display Behaviour

We’ve added more control over how the initial message appears in your chat widget.

What’s New:

• New toggle in chatbot embed settings to choose when the initial message is shown:

  • Only on first page load (default)

  • On every page load

• This setting is saved per bot and applied across all embedded instances.

• Gives users flexibility to balance visibility with user experience.

Now you can customize how often your visitors see the welcome message, depending on your engagement strategy.

Moved Email-Related Tasks to High-Priority Queue

We’ve restructured our task queue system to ensure faster delivery of time-sensitive email actions.

What’s New:

• All email-related tasks, including OTPs and notifications, are now routed through a high-priority queue.

• Ensures quicker processing and reduced delay in critical communications.

• Default task queue now reserved for non-urgent background jobs.

This change improves reliability and responsiveness for essential user interactions like login verifications and email alerts.

✅ Fixes & UI Enhancements

Fixed Error Occurring When Calling Knowledge Base Function

Resolved a backend bug that caused intermittent failures when invoking knowledge base functions.

Fix Highlights:

• Improved error handling and fallback logic.

• Enhanced logging to better track usage and failure points.

Conditionally Show/Hide "Edit" Button Based on User Permissions

UI now respects role-based access control more strictly:

• “Edit” button in drop downs only shows for users with edit permissions.

• Cleaner and safer UI experience for restricted users.

Font Mismatch on Apple Devices in AI Agent Chat Widget

Standardised font rendering across all platforms:

• Fixed inconsistent font display on Safari and other Apple-based devices.

• Ensures a more uniform and professional appearance.

Fixed Chat Widget Popup Message Position (Left-Side Placement)

• Resolved positioning issues when the chat widget is docked on the left side.

• Popup messages now appear properly aligned with widget boundaries.

Cooldown Message Not Displayed for WhatsApp/Meta Rate Limiting

Users now receive feedback when soft rate limits are hit:

• Cooldown messages for Meta/WhatsApp integrations now appear consistently.

• Helps reduce confusion around failed message deliveries.

Character Limit for Chatbot Query

To ensure stability and performance, we’ve introduced a character limit for chatbot queries.

What’s New:

• A 10,000 character limit is now enforced for all users not using their own API keys.

• This applies based on the primary model selected for the chatbot (e.g., OpenAI, Claude, Grok).

This update helps maintain system reliability and prevents overload from excessively long inputs.

UI Refinements

Vision Section Image Stretching:

• Images in the Vision section now maintain proper aspect ratios.

  

Popup Viewer for Chat Images:

• Clickable chat images open in a dedicated viewer with better scaling.

  

File UI Update:

• Refined file preview elements, improved upload feedback and styling.

  

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

• Soft Rate Limit with Cool Down & Custom Message:

  To prevent spam or abuse, we’ve introduced a Soft Rate Limiting System with customizable settings and cooldown behavior.

What’s New:

• Configurable Limits: Set message caps per hour or day.

• Cool Down Periods: Temporarily block users from sending messages once the limit is exceeded.

• Custom Messages: Display personalized messages explaining the reason and duration of the cooldown.

• Admin Controls: Configure all parameters via admin settings.

• Logging: All rate-limit triggers are logged with user ID and usage details.

Example:
If a user sends 100 messages in an hour, they will enter a 30-minute cooldown and see:
“You have exceeded the message limit. Please wait 30 minutes before sending more messages.”

• Agency/Sub-account Analytics:
  Introduced analytics tracking for agency and sub-account activity to give better visibility into usage patterns and engagement metrics.

✅ Fixes & UI Enhancements:

• Display Label for outstanding_subscription_fee
  Replaced internal value references with user-friendly labels for outstanding subscription fee details to improve readability.

• Human Handoff UI Refinements

  Improved the Human Handoff settings UI to better match mockups and enhance usability:

  • Reduced padding and fixed alignments

  • Added missing "Unavailable" label

  • Centered "Available 24 hours" text

  • Made time dropdown searchable with fixed height

  • Expanded rows and dropdowns per design

  • Fixed layout issues on validation errors

• Display Error Message on AI Agent Training Failure
  We've improved the feedback users receive after AI Agent training runs.

  What's New:

  • When training completes with partial errors, a detailed summary is displayed.

  • Example: “Training failed for 1 and successful for 44 data file(s). All QnA trained successfully.”

  • Total token usage is also shown for transparency.

  This helps users quickly understand training outcomes and address specific file issues without confusion.

• AI Agent Settings Form Improvements

  • Made "Name" an optional field.

  • Automatically scrolls to the field with validation error upon form submission.

• Inaccurate Character Count in Scraped URLs & “Select All” Issue
  Fixed discrepancies in character counts for scraped URLs and addressed issues where “Select All” was not applying properly across paginated entries.

🧠 Knowledge Base Improvements:

• Scraped URLs Tab – Pagination Upgrade
  Replaced infinite scroll with numbered pagination for better performance and navigation in the Scraped URLs tab.

• Files Section Enhancements
  Added powerful new capabilities in the Knowledge Base > Files section:

  • Filter files by Trained/Untrained status

  • Search by file name

  • Pagination support

  • Multi-select actions for bulk operations

🔁 AI Agent & Marketplace Updates:

• Add Option to Delete AI Agents from Marketplace
  Users can now remove AI Agents from the public marketplace with a single click.

• Scrollable Dropdown with Search in AI Agent Cloning Modal
  Improved usability of the sub-account selection dropdown when cloning agents by adding a fixed height, scroll support, and search functionality.
  

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

• MasterChat v2 (Vision + Documents)
  We’ve introduced new capabilities to enhance AI agent interactions with visual and document data.

  • Upload Images & Documents:
    Users can now upload images and documents while querying the AI agent and receive responses based on those files.

    

  • Enable Vision Features:
    To activate this functionality, users must check the “Enable Vision Features” option in their AI agent’s settings tab.

    

  • SaaS Configurator Support:
    Agencies can set rebilling prices for vision-related usage via the Free Trial Settings tab in the SaaS Configurator.

    • It automatically calculates profit to simplify billing management.

      

• Pagination Added to Stripe Payments tab in Billing Settings
  To enhance usability and performance, we’ve introduced pagination to the Stripe income and payment history tables in Account & Billing Settings.

  • Improved Performance:
    Previously, all transactions were loaded on a single page, which caused slow load times and a cluttered view when dealing with a large volume of entries. Pagination now ensures smoother navigation.

  • User-Controlled Entries Per Page:
    Users can now choose how many records to display per page — with options including 10, 25, 50, and 100 — allowing for a more tailored and efficient browsing experience.

  • Where to Find It:
    You can access this update in Account & Billing Settings → Payments Tab
    Go to Stripe Transactions

• Upgraded Custom Actions (Functions)
  We’ve enhanced the custom actions framework by introducing support for static parameters. Previously, all parameters had to be dynamically extracted from conversations at runtime. With this update, you can now define static values directly in your functions — perfect for use cases where certain inputs never change. This adds more flexibility and control for advanced automation workflows.

• Chatbot Embed Position Toggle – Bottom Left or Right
  You can now customize the position of the embedded chatbot widget on your website. A new toggle has been added to the embed settings, allowing you to choose whether the AI Agent appears in the bottom left or bottom right corner. This gives you greater flexibility to match your site’s layout and design preferences.

  

🛠️ Fixes & Improvements

• Optimized KnowledgeBase Chunking Strategy
  We've improved the chunking logic to ensure your content is no longer cut off or missed during processing. This advanced strategy enhances semantic search accuracy, leading to more relevant and complete responses from your AI agents.

• Fixed Chat Widget Input Field UI Bug in code – Unwanted Top Border on Background Change
  Resolved a UI issue where an unintended top border would appear in the Chat Widget input field when a background color was applied. This fix ensures a clean, consistent appearance regardless of styling.

• Cloning AI Agent from Agency View Should Trigger Sub-account Selection Modal
  Previously, cloning an AI Agent from the Agency view would immediately duplicate the agent without asking for the destination sub-account — unlike the behavior within an AI Agent view, where a modal prompts sub-account selection. We've updated the Agency view to match this intended behavior, ensuring users are always prompted to choose the target sub-account during cloning. This provides more control and prevents agents from being cloned into the wrong account.

  

• Human Handoff Responses Not Delivering on Meta Apps
  Resolved an issue where live human handoff messages were not being delivered correctly on Meta Apps(i.e WhatsApp, Facebook, Instagram).

• Scraped URLs Count Mismatch & Duplicate Status Messages
  We've fixed an issue where the number of URLs shown in the Scraped URLs tab did not match the count displayed after scraping. Additionally, success and error messages were being duplicated for a single event. This update ensures that the URL count reflects the actual number scraped and that status messages are displayed only once per event — improving clarity and reliability in your knowledge base management.

• Incorrect Time Conversion in Leads Table for Certain Timezones (e.g., US Timezones)
  Time conversions in the Leads table are now accurate across all timezones, including various US timezones. This fix ensures consistent lead tracking and reporting regardless of user location.

• Update Wallet Usage Card on Dashboard – Rename to "Wallet Balance" and Display Available Balance
  The Wallet Usage card on your dashboard has been renamed to "Wallet Balance" and now displays the actual available balance for clearer and more intuitive wallet management.

• Search Filter in "Scraped URLs" Tab Displays Irrelevant Results
  Improved the search filter in the Scraped URLs tab to return more relevant results, making it easier to find specific URLs and manage your data sources efficiently.

• Functions Not Being Duplicated When Agent is Cloned
  Fixed an issue where custom functions were not being duplicated when cloning AI agents. Now, all functions will properly copy over, saving you setup time.

• Search Bar Not Functioning in Dashboard Sub-Accounts
  The search bar in dashboard sub-accounts is now fully functional, helping you quickly find agents, users, or settings across your managed accounts.

• Initial Popup Messages Not Displaying on Websites (Except in Incognito Mode)
  Fixed an issue where initial popup messages were not showing up on websites in normal browsing mode. This fix ensures all users see the popup messages as expected.

• SaaS Configurator Updates

  • Hide Deprecated Models for New Subscriptions:
    Only non-deprecated models are now shown when creating new subscriptions. Legacy users can still access older models.

  • Updated Model Titles:
    Improved clarity for OpenAI and alternative models, including version tags and deprecation labels.

  • User Avatar Icon:
    Users can now upload a custom avatar for chat participants instead of using the default grey icon.

  • Improved Cost Messaging:
    When an API key is connected, model pricing shows API costs; otherwise, default costs display.

• AI Summary in Conversations Tab Not Working for Some Bots
  Fixed a bug where the AI-generated summary in the Conversations tab was not displaying for some bots. Summaries should now appear consistently across all supported agents.

• Updated Auto-Scraper to Count Only Additional Characters for AI Agent Training
  The auto-scraper now correctly counts only the additional characters added during AI agent training, ensuring more accurate usage metrics and billing.

• Fixed Inaccurate Character Count Display in Scraped URLs Tab
  Corrected the character count display in the Scraped URLs tab so it accurately reflects the total content size, helping you better manage your knowledge base limits.

• Duplicate Handoff Trigger Message and Misattributed User Messages in Conversation Tab and Widget
  Resolved an issue causing duplicate handoff trigger messages and misattributed user messages in the Conversation tab and chat widget, leading to clearer, more reliable conversation flows.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

Captcha Added to Signup Page
To improve platform security, we've added a Captcha to the signup flow to help block spam signups and bots.

• For agencies: You can enable this feature for all your sub-accounts by going to your Agency Profile. Scroll to the bottom and check the box labeled "Enable reCAPTCHA Verification" — this will make reCAPTCHA verification mandatory during signup for all sub-accounts.

Backup Models for OpenAI ChatGPT
We've added support for backup models to ensure smoother and more reliable AI agent performance during outages or high-traffic periods.
Now, you can select a backup model in your AI agent settings. If the primary model becomes unavailable, your agent will automatically switch to the selected backup — ensuring it keeps responding without interruptions.

The following backup models have been added:

• GROK-2-1212

  • Backup Cost: $0.020/message

• CLAUDE-3.5-Haiku

  • Backup Cost: $0.010/message

• CLAUDE-3.7-Sonnet

  • Backup Cost: $0.030/message

This new feature helps keep your AI agents online and responsive even when your primary model is down.

"Daily Support Call" Menu Item for Agency Users
A new menu item has been added to the sidebar for agency users to easily access the Daily Support Call link with one click.

Avatars in ChatWidget & Conversations Tab
Both bot and user messages now include avatars in the ChatWidget and the Conversations Tab for a more personalised and intuitive experience.

Human Handoff: Configurable Duration
You can now configure the handoff delay for live agents, giving you more control over when human support steps in.
In addition to setting the delay time, you can now:

• Specify days of the week when human handoff is allowed.

• Set specific time ranges during which handoff can be triggered.

Outside of the defined days and times, the handoff will not be triggered — and the AI agent will continue responding automatically. This ensures handoffs only happen when your team is available.

Upload Custom Logo for Masterchat in Agency Styling
Agencies can now upload a custom logo specifically for Masterchat under the Customization tab. This allows you to fully white-label the Masterchat experience with your agency’s branding, delivering a more professional and personalised interface for your team and clients.

Knowledge Base Testing Tab
A new "Knowledge Base Testing" tab has been added for all users, allowing you to interactively test and validate AI agent responses based on KB content.

🛠️ Fixes & Improvements

• Dark Background Behind Estimated Token Count in Base Prompt Section
  Fixed the issue of unreadable text by applying a dark solid background only behind the estimated token count in the base prompt section. The rest of the prompt area remains unchanged — this fix improves visibility without affecting the overall design.

• Marketplace Upload – Character Limit Fix
  Resolved character count restriction that was preventing certain AI agents from being uploaded to the marketplace.

• Webhook URL Limit Increased
  The allowed character length for webhook URLs has been increased to 300 characters, enabling better flexibility.

• My AI Agent UI Enhancements
  We’ve increased the size of the eyeball icon and the three-dot menu, and re-added a circular background for better visibility and alignment.

  

• Mobile Chat Widget Style Improvements
  We’ve fixed an issue where the initial popup messages on the chat widget were getting cut off on mobile view. This update ensures that popup messages display correctly and the widget behaves more responsively across all devices.

• Fix: AI Agent Card Close Issue
  Resolved a bug where the embedded AI agent card in the "AI Agent" tab wouldn’t close as expected.

• Total Character Count for Scraped URLs
  The total character count in the Scraped URLs tab now accurately reflects the content of all selected URLs — including when you select "Select All". This ensures you get a precise count of the total characters being used in your knowledge base.

  

• Integration Tab UX – Skeleton Screens
  We’ve replaced the generic "Loading" text with sleek skeleton loaders in the Integrations tab for a smoother and more modern UX.

• Minor UI Fixes

  • Fixed small formatting issue on the registration page.

  • Updated image in White-label documentation to remove “Stammer.”

  • Standardised height across all subscription package cards for a more consistent layout.

  ---

  New Live Support Options

  We’re excited to announce a bunch of new live support options to help you out!

  1. (NEW) ☎️ Daily Support Calls
     Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
     Link to join: https://stammer.ai/daily-support-call

  2. (NEW) 🚀 Thursday Agency Accelerator
     If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
     Link to join: https://stammer.ai/accelerator

Display Iframe in Chat Widget:
Improving rendering performance and user experience, by displaying iframe as HTML elements instead of being rendered at runtime.

Added Tooltip Information for Daily Scraper in Knowledge Base Tab:
Tooltip information for the daily scraper in the Knowledge Base tab has been updated for better clarity and guidance.

Master Chat Improvements:
Filters and search functionality have been added to the Master Chat Feedback feature, enhancing usability and providing more detailed insights.

• Redesigning the AI Agents Dashboard Experience:
  Revamped the AI agents' dashboard to deliver a more intuitive and streamlined experience. The updated design offers clearer, more organised information, with the added functionality of tagging agents and a powerful filtering system for easily sorting and viewing agents based on specific criteria.

• Embed Code & Meta Integrations Updates:
  The embed code has been moved back to the integrations tab, with separate tabs and separate permissions now available for Embed Code and Meta Integrations for better organisation.

  

• Text Update for Clarity - "Initial Messages" to "Initial Messages/Popup Messages":
  The terminology has been updated from 'Initial Messages' to 'Initial Messages/Popup Messages' to clarify that popup messages are considered the same as initial messages.

  

• Added Conversation UUID in "Copy Share Link" and Search Filter by ID:
  A unique conversation UUID has been added to the "Copy Share Link" feature, with an option to filter search results based on the conversation ID.

  

• Popup Bubbles in Chat Window Adapt Based on Content:
  Popup bubbles in the chat window will now adjust their size based on the content, improving the user experience and readability.

• Implemented TokenCounter in Base System Prompt Area:
  The TokenCounter has been integrated into the base system prompt area to provide more efficient token counting and management.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

Display Character Count for Selected URLs in Scraped URLs Tab in Knowledge Base

• The "Scraped URLs" tab now correctly displays the character count required to train the bot for selected URLs.

Total Characters Count Not Displaying for Some Users in Storage Table

• Fixed the issue where the total character count was not displaying for some users in the Storage Table. The character count is now correctly shown for all users.

New Live Support Options

We’re excited to announce a bunch of new live support options to help you out!

1. (NEW) ☎️ Daily Support Calls
   Need help getting started or figuring something out? Join our new daily 30-minute calls with co-founders Ross & Tim. Available Monday to Friday from 11:30am–12:00pm PST.
   Link to join: https://stammer.ai/daily-support-call

2. (NEW) 🚀 Thursday Agency Accelerator
   If you’re running an agency, don’t miss this! Every Thursday at 11:00am PST, we walk through client growth strategies, scaling tips, and proven workflows.
   Link to join: https://stammer.ai/accelerator

• AI Agent Analytics View: Introduced a new analytics view for AI Agents to track and analyse usage more effectively.

  

• Human Handoff Trigger for API conversations (with Live Chat): A new Payload URL feature has been implemented to handle human responses when triggering an AI Agent handoff via API. This allows users to specify where the human agent responses should be sent once the handoff occurs.

  Steps to Use the Human Handoff Feature:

  1. Trigger the Handoff: When you trigger the AI Agent handoff, ensure that the request includes the recipient_url in the body of the API request. This is crucial for the handoff process to trigger properly.

  2. Include recipient_url in the Request Body:

     • The recipient_url should be a valid URL (such as a web hook URL) where the human agent's responses will be sent once the handoff is triggered.

     • Example request body:

       jsonCopy{
         "chatbot_uuid": "12345678-1234-5678-1234-567812345678",
       
           "query": "Your message/string here.",
       
           "user_key": "Your message/string here.",
       
           "recipent_url": "https://example.com",
       
           "custom_base_system_prompt": "Your message/string here."
       }
       

  3. What is recipient_url?

     • The recipient_url is a URL endpoint where the human agent's responses will be sent after the AI Agent has completed its process and handed off the conversation. This URL could be a web hook, a server endpoint, or any other URL that can handle and process the incoming response.

     • For example, you can set up a web hook listener that will receive the human agent responses once the handoff happens.

  4. Ensure Valid URL: If the recipient_url is not included or is invalid, the handoff will not trigger, and the human response will not be delivered.

  5. Receive the Human Agent Response: Once the handoff is triggered successfully, the human agent will respond, and the response will be sent to the recipient_url provided in the API request.

• Added New Add-Ons Table for Upselling Subscriptions: Added a new table to help upsell subscriptions with more visibility and ease.

• Added Image Rendering for WhatsApp and Instagram: Enabled image rendering support for WhatsApp and Instagram within the platform.

• Knowledge Base Scraping: Added a check to enable scraping of Google Documents in the Knowledge Base.

  

• Added "View History" Button for Web hook History on Dashboard: A new "View History" button is added which allows users to see their Web hook history directly from the dashboard.

• Increased Web hook URL Length Limit: The Web hook URL length limit has been increased to 300 characters to support longer URLs.

• Add "Copy Sub Account Info" functionality: Users can now copy sub-account information easily, similar to the existing "Copy Support Info" functionality.

  

🐞 Bug Fixes & Improvements:

• Fixed Duplicate Sentence and Spelling Errors in White Label API Documentation: Removed duplicate sentences and corrected a spelling error in the API documentation for improved clarity.

• Prevent Concurrent Wallet Transactions: Implemented measures to prevent concurrent wallet transactions that could cause conflicts or errors.

• Debug Translated Emails: Investigated and resolved an issue where translated emails were not being sent in the correct language and defaulted to English.

• Free Trial User AI Agent Visibility Issue: Fixed the issue where free trial users were missing the AI Agent after subscription cancellation.

• Minor Fixes in MasterChat: Several minor improvements and bug fixes were made to enhance the functionality of MasterChat.

• Next Chat Opening After Deleting Conversation in 2.0: Resolved the issue where the next chat was not opening properly after deleting a chat in the 2.0 conversation view.

• Re-added Streaming Check to AI Agents Settings Tab: Restored the streaming check feature within the AI Agents settings tab.

• Spelling Error Fix in Documentation: Corrected spelling errors in the product documentation to ensure accuracy.

• Sync Skeleton with Dashboard View: Ensured that the skeleton view is synced properly with the dashboard for consistency across the platform.

• New Agents Showing Default Prompt Instead of Agency Default Prompt: Fixed the issue where new agents created for agencies or sub-accounts were showing the default prompt instead of the agency’s custom prompt.

• Leads Collection & Data Capturing Issue in Leads 2.0: Resolved an issue with leads collection and data capturing within Leads 2.0.

• Added User Key in Leads Data: Added a user key field to the leads data to improve lead tracking and management.

Note: MasterChat has been released for the Full SaaS Mode users 🚀.

• API Endpoint Expansion:
  We have introduced several new API endpoints to enhance the functionality and flexibility of our platform. The following new endpoints are now available:

  • AI Agent Endpoints:

    • Create AI Agent: Allows you to create a new AI agent.

    • Retrieve AI Agent: Fetches details of a specific AI agent.

    • Update AI Agent: Updates the configuration and settings of an existing AI agent.

    • Delete AI Agent: Deletes a specified AI agent from the system.

  • Knowledge Base Endpoints:

    • Add AI Agent Q/A: Adds new question and answer pairs to the AI agent’s knowledge base.

    • Add URLs: Allows you to add URLs to the AI agent's knowledge base for reference.

    • Add PDF File: Uploads a PDF file to the knowledge base for the AI agent to process.

  • Conversations Endpoints:

    • AI Agent Conversations: Retrieves a list of conversations for a specific AI agent.

    • Retrieve Conversation: Fetches the details of a specific conversation.

    • Delete Conversation: Deletes a specific conversation from the system.

  • Sub-Account Endpoints:

    • Retrieve Sub-Account: Fetches details of a specific sub-account.

    • Update Sub-Account: Updates information and settings for a sub-account.

    • Delete Sub-Account: Deletes a sub-account from the system.

  For detailed information on how to use these new endpoints, please refer to the API Documentation.

• Size Options for Chat Widget in Appearance Tab:
  Users can now select from three size options (Small, Medium, Large) for the chat widget in the Appearance tab, offering improved customisation for your website’s design.

  

• Appointment Scheduling Web hook – Included Collected Leads Data:
  The Appointment Scheduling Web hook now includes collected leads data, providing more context and information with each appointment request.

🔍 Improvements:

• Google Sign-In Button Update:
  The "Sign in with Google" button has been updated to the official Google branded asset, providing a more consistent and familiar user experience.

• Sub-Account Status for Direct Billing:
  Added a new status indicator for directly billed tags in sub-accounts, giving more visibility and clarity on billing statuses.

• Bypass Sub-Account Billing in SaaS Configurator:
  Added a new toggle inside the SaaS configurator to Bypass Sub-Account Billing. This will allow you to sell a subscription package for a fixed monthly amount, but the client will not be charged for usage (your agency is still charged from your wallet, or if you’re using your own API key, that will get charged). If enabled the client sub-account will not have a wallet balance.

  

• Pagination in Scraped URL Table:
  Pagination has been added to the Scraped URL table, improving performance and usability when viewing large sets of scraped URLs.

🐞 Bug Fixes:

• Recurring Credits Display Issue (Monthly Not Saving Correctly):
  Fixed a bug where recurring credits were incorrectly displayed as one-time credits, ensuring monthly recurring credits are now saved and displayed correctly.

• Test Function in Scheduling Not Displaying All Available Slots:
  The issue with the Test Function not showing all available time slots for scheduling has been resolved, ensuring users can now see the complete range of available dates and times.

• Incorrect Date and Contact Info Storage in Collected Leads:
  Fixed the issue where dates in collected leads were mixed up and contact information provided by the bot was being stored incorrectly. Data is now properly stored and formatted.

• Crawled URL – Select All Not Working Properly:
  Fixed an issue where the "Select All" feature for crawled URLs was not functioning as expected, improving efficiency when selecting multiple entries.

• Redirect Links for White Label Incorrectly Leading to Login Page:
  Resolved the issue where white-label redirect links were incorrectly routing users to the login page, ensuring correct redirection after actions.

Note: Masterchat will be releasing in our next release 🚀.

As soon as the user has answered all of the questions, the AI agent will immediately extract their responses and send the data to the notification email/ webhook.

Based on the values you enter, the system will generate a prompt that you can copy and paste into your base system prompt.

Replace the Custom_Variables with the data you want to extract and add the following prompt to your base system prompt after AI agent data collection instructions:

## Data Collection Instructions:
After you have collected all of the data from the user: {Custom_Variable1, Custom_Variable2...}, call the send_notification_about_lead_collection() function.

How to Implement:

Step 1 - Write out your base system prompt to instruct the AI agent what to do, how to act and respond to users. This includes instructions to ask the user questions like “Ask the user to provide their email address”.

Step 2 - Go to the Leads tab in your AI agent and add the data values of contact information the AI agent should extract from the conversation like ‘email address’. Enter in an email or webhook to send the data to.

Step 3 - Copy the prompt instructions into your base system prompt. “After you have collected all of the data from the user: {email}, call the send_notification_about_lead_collection() function.”

The AI agent will now be able to collect the data and send it after the user has answered all of the questions. If the user only fills out some of the data fields whatever data is available will be extracted when the conversation ends or after 1 hour of no activity.

Learn More:

https://docs.stammer.ai/stammer.ai-docs/ai-features/leads-2.0

• Implement Redesign of Knowledge Base Section: The Knowledge Base section has been redesigned for a cleaner, more organised look, improving ease of navigation and access to resources.

  

• Bot Gets Stuck When Receiving Timezone Response During Meeting Scheduling: We've resolved an issue where the bot would get stuck when receiving a timezone response during meeting scheduling, improving scheduling flow reliability.

• Auto-send Verification Email on First Login: To enhance security, a verification email is now automatically sent to users upon their first login.

• Format Wallet Quota Display: The wallet quota display has been formatted for clarity, making it easier to track usage and available funds.

• Allow Users to Start Free Trial After Cancellation in Warning Period: Users who are in the warning period after cancellation can now restart their free trial, providing more flexibility with trial management.

• Update Error Messages for Better Clarity and Accuracy: Error messages have been updated to be clearer and more accurate, ensuring users understand issues better and can take appropriate action.

• Re-Implement Legacy Conversations AI Summary Feature in Live Chat: We’ve brought back the AI-generated summary feature for legacy conversations! After it was temporarily missing in the previous release, we’ve now re-implemented it, ensuring that both new and past conversations are enhanced with AI summaries for easier review and quicker responses.

• Add Filters to Storage Transactions: We've added filters to storage transactions, making it easier to find and manage specific records.

• Display Chatbot/Human Response Labels Under Messages in Conversation 2.0 Tab: Clear labels now appear under messages in the Conversation 2.0 tab, distinguishing between chatbot and human responses for better clarity.

  

• Implement Filter Functionality for Exporting Conversations in Live Chat: You can now filter conversations when exporting from Live Chat, making it simpler to retrieve specific data.

• Display Proper Exception Error When Card Update Fails: Card update failures now display the correct exception error, making troubleshooting easier for users.

🚀 New Features:

Lead 2.0 – A Major Update for Enhanced Lead Collection 📊

Leads 2.0 represents a major overhaul of the lead collection system, designed to offer greater flexibility, organisation, and customisation for how you manage your leads. Here’s a breakdown of the key changes:

1. Streamlined Lead Collection Process

• No Separate Prompt for Leads: In Leads 2.0, there is no longer a separate prompt for collecting leads. Instead, you now define the lead collection criteria directly within the base prompt of your AI agent.

  • What does this mean? You can define what specific data you want to extract—such as email addresses, names, appointment times, or any other custom details—right in the base prompt.

  • The data will automatically be collected when the conversation ends or after 1 hour of inactivity. This creates a more seamless and integrated lead collection flow.

  • If you add any new data fields, remember to update your AI agent's base system prompt to ensure the system asks the user for the required information.

2. Redesign of AI Agents Settings Tab

We’ve revamped the AI Agents Settings tab to offer a more intuitive and user-friendly experience, giving you better control over the settings and configurations of your agents.

🔍 Improvements & Enhancements:

• Auto-save Human Handoff Keywords and Email in Conversation Tab: Human handoff keywords and emails are now automatically saved in the conversation tab, ensuring all necessary information is stored for easy reference during handoffs.

🛠 Bug Fixes:

• Cancelled Subscription Prevents Access to Free Trial: Users who had canceled their subscriptions were previously unable to access free trials. This issue has been resolved, so you can now freely access the trial even after canceling.

• Fix Formatting Issues in Live Chat UI: Several formatting inconsistencies in the live chat interface have been addressed for a cleaner, more consistent look.

• Image Rendering Discrepancy Between Dashboard and Meta: Fixed discrepancies in how images were rendered between the dashboard and Meta, providing a consistent experience across both platforms.

• Incorrect Message Date in Live Chat: We’ve corrected an issue where the dates of messages were incorrectly displayed in the live chat interface.

These updates bring significant changes, especially with the rollout of Leads 2.0, providing more flexibility in lead collection and a smoother user experience overall. Enjoy the new features and improvements! 🎉

• Eye View for Previewing Scrapped Files: We’re excited to introduce an intuitive "eye view" feature, giving users a sneak peek of scrapped files just like link previews. This makes it easier to see content at a glance before taking any action!

  

• Toggle Option for White Label Signups: You now have more flexibility! With a simple toggle in the settings, users can turn off white label signups when needed—making customization easier than ever.

  

🔍 Improvements & Enhancements

• Increased URL Length for Webhook Lead Collection: The URL length for webhook lead collection has been increased to 300 characters, allowing for more robust and detailed data collection!

🛠 Bug Fixes

• Nginx Configuration Fix: The issue of automatic Nginx configuration removal has been resolved, ensuring that custom configurations remain intact and uninterrupted.

💬 Live Chat Updates

• Human Handoff via API: You can now request a human handoff through the API, making it easier for users to escalate conversations quickly when needed.

  

These updates bring even more flexibility and clarity to your experience, with new features, important fixes, and improvements designed to make everything work seamlessly. Enjoy! 🎉

🚀 Key Features:

1. Human Takeover from AI Agent:

   • Human agents can now take over a conversation from the AI agent at any time, either through the Stammer dashboard or directly via the native chat application.

   • When the AI agent encounters a situation it cannot handle, the conversation can be seamlessly handed off to a human agent.

     

2. Live Chat Human Handoff Automation:

   • Human Handoff can be automatically triggered based on specific criteria defined by the user (e.g., certain words or phrases).

   • Once triggered, the AI agent stops responding and sends a notification to a human agent to take over the conversation.

     

3. Trigger Settings:

   • Customizable words/phrases can be configured to detect when the AI agent should stop responding and trigger the handoff.

   • This helps ensure that complex or sensitive conversations are escalated to a human agent without delays.

     

4. Handoff Response Customization:

   • Customize the message the AI agent sends before transferring the conversation to a human.

   • Example: "It seems I’m unable to provide a satisfactory answer. A human agent will take over shortly."

     

5. Notification Settings:

   • Configure the notification recipients when a conversation requires human attention.

   • Notifications can be sent to multiple emails or through a webhook for integration with other systems.

     

6. Platform Integration:

   • Human Handoff can be enabled for specific channels that your AI agent is connected to.

   • For Meta integrations (Facebook, Instagram, WhatsApp), Human Handoff automation will only work if you have an active connection to the platform.

     

7. Native Chat Integration for Human Handoff:

   • Human agents can take over conversations directly within integrated platforms like Facebook Messenger, Instagram DM, and WhatsApp.

   • Users can enable Human Handoff for specific platforms by toggling the settings on each platform.

   • When a handoff is triggered, agents can respond directly from the Stammer dashboard, providing a unified place to manage conversations.

   • Agents also have the option to continue the conversation via the native chat app, offering flexibility to manage support from multiple channels.

   • This integration ensures that agents can provide personalized, real-time support without needing to switch between different platforms.

Example Email Notification:

• Smarter Meeting Scheduling:

  • Meeting Descriptions in Scheduling Prompts: Added meeting descriptions to the scheduling prompts, so now you can also get meeting descriptions from users during the booking process.

  • Get Meeting Descriptions After Booking: You can also get meeting descriptions after a meeting is booked. Simply uncomment the two highlighted lines in the default scheduling prompt, and you're all set.

  

• Advanced Image Capabilities:

  • Scraping & Crawling Images: Your bots now have the ability to scrape and crawl images from the knowledge base, adding visual elements to the information they retrieve.

  • Image Sharing in Responses: The bots can now share relevant images directly in responses, allowing for richer, more engaging conversations with users. This helps to provide more context, improve user understanding, and make the interaction more dynamic.

• Streamlined Integration Dashboard:

  • Redesigned Interface: The layout has been redesigned to make managing integrations more intuitive and user-friendly.

  • Embedded AI Agent & Chat Widget Settings: These settings have been relocated to the Integration Tab, which may make it easier for users to access and manage both integrations and AI-related features in one place.

  

• Advanced Scraper Enhancements:

  • Wait Time for Browser: Introduced a new functionality where the browser waits for a specified duration before starting the crawl. This allows for content to load properly, especially on dynamic websites.

  • Scroll to Bottom Check: Added a "scroll to bottom" feature for websites that load content as the user scrolls. This ensures the entire page loads before the scraper begins collecting data.

  

🔍 Enhancements

• Optimised Performance: We've improved the speed across all tabs and pages, ensuring a smoother and more responsive experience for users.

• Enhanced Scheduling Algorithm: The scheduling algorithm has been optimised for better accuracy and efficiency, resulting in more reliable and seamless booking experiences.

• Enhanced Storage Management: Implemented pagination for storage and AI Agent subscriptions in Billing Usage & Settings tab, making it easier to navigate large datasets and manage information efficiently.

  

🐞 Fixed Issues

• Human Handoff: The system now stores and displays triggered messages during human handoff cases, ensuring a seamless transition and continuity in conversations.

• Brazilian Portuguese Translations: The Brazilian Portuguese translations have been updated for greater accuracy and clarity.

• API Key Management: Adjusted the API key naming rules to prevent 400 errors, ensuring smoother functionality and integration.

• Back Button on Custom Function Page: Introduced a back button on the custom function page to enhance user navigation.
  

  

• Free Trial & Wallet Balance Inputs in SaaS Configurator: Added input fields for managing free trial durations and wallet balances directly in the SaaS configurator.
  

  

🔍 Enhancements

• Consistent Markup Language Rendering: Ensured consistent rendering of markup language across all integrations, improving UI stability and readability.

🛠 Fixes

• Spanish Translations in Email: Added missing Spanish translations in email templates for a more inclusive user experience.

• WhatsApp Display Name Handling: Improved error handling for unapproved WhatsApp display names to prevent disruptions in user messaging.

• Consistent Popup Text Above Chat Bubble Widget: The popup text above the chat bubble widget now seamlessly inherits the font size from the Appearance tab, offering a polished and cohesive design.

• Improved French Translation: Updates to French translations have ensured a more intuitive and enjoyable experience for our French-speaking users.

🛠 Fixes

• Resolved Symbol Display Issue: Special symbols in user responses now render correctly in the Conversations tab without being converted into HTML entities.

• Duplicate Sub-Account Entries Eliminated: Duplicate sub-account entries in search results have been addressed, ensuring clean and accurate results.

• Character Validation for Downgrades: Added a character validation check to ensure hassle-free package downgrades with no interruptions or errors.

• Improved SSL Certificate Renewal: The renewal process has been streamlined to ensure smooth DNS uploading and hassle-free management of certificate expirations.

• Subscription Webhook Alerts for Agencies: Agencies now have access to subscription webhook alerts, enabling them to receive notifications for failed user subscriptions or payments.
  

  

• Stripe 3D Authentication Support: Enabled wallet credit recharge functionality with support for Stripe’s 3D Authentication.

🔍 Enhancements

• Enhanced Notifications for URL Crawling and Training: Improved the notification messages displayed after a website is crawled or trained, now including details on the number of successful and failed URLs.

• Manual Search for URLs in Website Scraper: Added functionality for manual URL searches within the website scraper, providing greater control and flexibility.
  

  

🛠 Fixes

• Improved Bot Scheduling: Enhanced the bot scheduling process, ensuring smoother functionality and improved performance.

• Bot Responses in Conversations Tab: Fixed a bug where HTML responses were not rendering correctly in the conversations tab.

• Pagination Issues for Selected URLs: Fixed a problem where selected URLs weren’t preserved across pages, improving consistency.

• Meta Integration on Free Trial End: Disabled Meta integrations automatically for sub-accounts after the free trial ends.

• Website Launch Error Messaging: Enhanced error messages for failed website launches caused by DNS issues for better troubleshooting.

• Missing Translations: Added missing translations for Japanese and Hungarian, ensuring a seamless experience for native users.
  

• New Language: Added Hungarian language to the app, enhancing accessibility for Hungarian-speaking users.
  

  

• New Currency: Added Swedish SEK as a currency option to the platform, enabling broader financial compatibility.
  

  

• Nested Crawling for Failed URLs: Enabled deeper crawling for failed URLs to improve data reliability and completeness.

• Conversation Summary and Lead Webhook Improvements: Fixed ordering issues in conversation summaries and webhook data to ensure accurate processing.

• Subscription Card Update: Displayed one-time setup fees directly on subscription cards for better transparency.

Bug Fixes:

• 500 Internal Server Error on Subscription Plan: Resolved errors encountered on 'Review Invoices' and 'Card Details' pages.

• Bot Widget Display Issue: Adjusted bot widget behavior to prevent shrinking on medium screens with available space.

• Instagram Account Linking Enhancement: Improved the account linking process to ensure each Instagram account is correctly associated with a single bot, enhancing reliability and user experience.

Critical Fixes:

• Bot Crawling Limits: Fixed a bug causing bots to crawl more links than the allowed maximum limit.

Enhancements:

• Dashboard Pagination for Sub-Accounts: Introduced pagination to the agency dashboard, simplifying the management of sub-accounts.

• Japanese Translation Update: Addressed missing Japanese translations across multiple modules to enhance localization.

• Spanish Translation Improvement: Enhanced Spanish translations in the chatbot dashboard, ensuring improved accuracy and a seamless user experience.

• Handoff on First Message: Resolved a bug preventing handoff to a human agent on the first user message.

• Extended Appointment Durations: Added new options for 90 and 120 minutes appointments, providing greater scheduling flexibility.

Security:

• Vulnerability Report Analysis: Implemented necessary changes to address identified vulnerabilities, including setting a password length limit of 72 characters for enhanced security.

• 2FA Authentication: Added two-factor authentication for enhanced security.

  

• Optimized URL List Performance: Improved loading times for datasets containing 2,000-3,000 URLs.

• SaaS Configurator Layout Fix: Enhanced layout usability in the Agency’s Subscription SaaS configurator.

• HTML Tags Handling: Prevented HTML tags from rendering as plain text in bot responses.

• Marketplace Chatbot Enablement: Introduced the ability to enable/disable chatbots from the marketplace interface.

Bug Fixes:

• One-Time Payment Fix: Addressed an issue causing one-time payments not to display correctly.

• Stripe Balance Overview Fix: Resolved text overflow due to missing horizontal scrolling in the Stripe balance transaction overview.

• Improved Human Handoff: Fixed a bug where the bot failed to hand over conversations to human support after triggering the handoff.

• WhatsApp Conversation Management: Resolved issues with conversations being combined for different end-users.

• Bot Customization Bug Fix: Ensured the bot interface scrolls correctly with the page.

• Scraping Issues Resolved: Fixed critical scraping-related problems to improve data reliability.

• Conversation Status Tracking: Resolved issue with conversations not being marked as completed, leading to empty webhooks.

• Customer Sub-Account Issue: Fixed error during agent launch in customer sub-accounts via the marketplace.

• Conversations AI Summary: Added for improved user insights.

• Starter Message Persistence: Ensures the starter message remains visible, even in ongoing conversations.

• Meta Integration Fixes: Resolved URL redirection issues.

• Translation Updates: Minor refinements for enhanced accuracy.

• Scheduling Improvements: Fixed TypeError related to schedule_appointment() and unexpected keyword argument 'location'.

• Conversations Flagging: Displays a flag if the AI summary has already been sent.

• Custom UI Enhancements: Improved the management interface.

• Italian Translations: Updated files for better localization.

• Info Cards on Whitelabel: Option to hide info cards.

• One-Time Extra Fee: Added feature for additional subscription charges.

Fixes and Updates:

• MRR Display Issue: Resolved issue causing MRR to display zero for some agencies.

• Wallet Balance: Fixed update sync from management via transaction records.

• Cancellation Data Check: Ensured proper handling for uncollectible, unpaid, and canceled subscriptions.

• Email Translations: Applied language translations to all emails per receiver preferences.

• Conversation Number Allotment: Prevented allocation of conversation numbers to empty or unlinked leads, resolving related errors.

Additional Improvements:

• Redesigned and simplified DNS configuration.

• HUF Currency Support: Added support for the Hungarian Forint (HUF) currency for agencies.

• Failed URL Visibility: Enhanced scraped URLs feature to show failed URLs for better debugging.

Improvements

• URL Sorting: Fixed issue with URLs not being sorted on page load.

• Agency Analytics Table: Updated the Agency Analytics table in the management dashboard for better clarity and usability.

• Question/ Answer Training: Quick fix to train question and answer pairs from ‘Training a Better Response”.

• Export Transactions: Improved export functionality to handle transactions over 1,000, enabling email delivery for large datasets.

• DNS Verification: Added a step to verify DNS settings before launching a website in Step 2.

Miscellaneous

• General Fixes and Improvements: Resolved various minor issues to improve overall stability and performance.

• Email Translation: Translated email notifications based on the receiver (Wallet, Storage, and AI Agents).

Fixes and Improvements

• Stripe Payment History:

  • Resolved issue where payments made to Stammer on production were not showing.

  • Improved UI for Stripe payment details.

  • Limited Stripe API hits during wallet recharges.

• Wallet and Auto-Scraper Enhancements:

  • Enabled storage recharge and balance checks in Auto-Scraper.

  • Synced wallet balance and transaction updates after auto-recharge.

  • Added a hook to ensure wallet balance and transactions are in sync.

• URL Scraping Issues:

  • Fixed multiple URL scraping issues, including URLs not being scraped.

• Sub-Account Management:

  • Fixed the billing tag display in sub-account cards.

  • Added pagination to the Sub-Accounts table on the dashboard for improved usability.

• Agency Default Language in Login/Signup: The Login and Signup pages now automatically display in the agency's default language, enhancing the user experience.

• Free Trial Config Access: Free Trial Configurations are now available in SaaS Configuration without needing a Card or Stripe connection.

• Separate Stripe Payments Tab: Stripe payments now have a dedicated tab for easier access and management.

• Wallet Transaction Export: You can now export wallet transaction histories for detailed record-keeping.

Enhancements:

• Increased Credit Limits in SaaS Configurator: Expanded credit limits within the configurator to better support client needs.

• Email Notifications for Successful Rebilling: Agencies now receive an email notification when storage subscriptions are successfully rebilled.

• Amount Range Filters in Stripe Payment History: Added range filters for better control and filtering of Stripe payment histories.

• Improved Labels and Placeholders: Updated labels to be more concise, added placeholders for user clarity, and included a tiny label below the Handoff switch for better understanding.

Fixes:

• URL Character Count Fix: Ensures accurate character counting of untrained URLs before training a chatbot.

• Filtered Conversation Deletion Issue: Fixed an issue where deleting a conversation removed the active filters unintentionally.

• Subaccount Signup Notification Fix: Adjusted notifications to ensure only a single email is sent to the agency when a subaccount signs up.

🚀 New Features

• Help Icon Behavior: Help icon now opens within the same tab for smoother navigation.

• Default Language Settings for Agencies: Agencies can now set a default language across their sub-accounts to streamline user experience and localization.

🔍 Enhancements

• Wallet Transaction Filters: Added search and sorting filters for easier management and review of wallet transactions.

• Usage Link on Dashboard: View usage metrics directly from a link on the dashboard card, allowing for faster insights into your app usage.

• Stripe Invoice/Transaction Redirect: Credit recharge transactions now include a direct link to the relevant Stripe invoice/transaction for seamless tracking.

🛠 Fixes

• Chatbot Marketplace Launch: Resolved a critical issue preventing some users from launching their chatbot in the marketplace.

• Chatbot Temperature Setting: Addressed a bug where temperature settings were not being saved when using the platform in French.

• Welcome Message Font Adjustments: Improved font size and family for the initial welcome message to enhance readability.

• Chat Widget Styling Improvements: Various design and styling fixes were made to enhance the visual consistency and appearance of the chat widget.

• Mobile Responsiveness: Optimized the chatbot share page for improved mobile usability.

• Chat Widget Icon Sizing: Adjusted icon sizes within the chat widget to improve visual clarity across devices.

New Currencies

• Japanese Yen

• Thai Baht

Updates & Fixes

• Scraped characters count update

• Add pagination for wallet transaction table

• Handle exception for embedded chatbot, when actual chatbot is being deleted.

• AI Agents stats (fix)

• Lighten up the billing account and agency detail pages in admin dashboard.

• Improve Card header responsiveness

• Update email fields to be a list of comma separated (All fields updated except conversation summaries which will be in the next release)

• Download invoices made to Stammer even after a plan is canceled.

• Empty conversations being sent in conversation summary

• Add conversation in Leads notification

• Show buying currency in Wallet recharge modal

• Chat widget styling issue resolved

• Fix: Character count in scraped urls

• Improve Downgrade Error Message for Exceeding Chatbot Limit

• Send and show google meet details when appointment is scheduled

  

This allows agencies to show/ remove the white label documentation from the left hand menu for sub-accounts. Agencies can hide the white label documentation and add their own custom documentation using the custom menu links.

2. Add google-meet link in scheduled event

Inside the scheduling webhook, the meeting link for Google meets will now be included. This will allow users to send meeting notifications from their CRM dynamically with all of the information collected from booking the meeting.

3. Bug fixes and improvements

We are constantly improving the platform and fixing edge cases.

4. Website Scraper Filters

You can now sort through all of the individual URLs that have been scraped to easily select the data you want to train your AI agent on.

🔥 1. Sub-account summary notifications

• Sub-Account users now have the ability to stay up to date with the status of their account.

• Track the total number of messages sent, total wallet balance and conversations within the client’s sub-account.

🔑 2. Subscription-id mandatory

• This is a quick update to creating new sub-accounts via the white label API. The subscription-id is mandatory if a paid account is being created using API.

📊 3. Agency Summary Notifications

• Agency users now have access to a summary of the statistics of your account on a daily, weekly or monthly basis.

• See how many messages your sub-accounts are sending, what your current wallet balance and storage balance is and more.

🛠️ 4. Directly billed sub-accounts using API

• Billed sub-accounts using API are now created directly.

🤖 5. Add overall used AI Agents for agency cards

• We now have a new row, for Overall Used AI Agents i.e. Agency Used + Sub-account allocated.

💵 6. Add "New Zealand Dollar" currency for Agencies

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

Improvements

1. Meta Integrations are out of Development:

   • Integrations for Facebook, Meta, Instagram are out of development and in Beta!

     

2. Fullscreen Mobile Chat Widget:

   • Introduced a (Updated & Improved) fullscreen chat widget for mobile users.

3. Dynamic Avatar Customization:

   • Added avatar customization that adapts based on the theme.

4. Improved Conversation Summary:

   • Enhanced the conversation summary by including a human-readable date and time format.

5. LLM Model Version Selector Relocation:

   • Relocated the LLM model field above the base system prompt for better organization.

     

6. Dynamic Naming for Appointment Invites:

   • Enabled dynamic naming for appointment invites on the calendar.

     

Bug Fixes

1. Public Conversation Page Layout Fix:

   • Fixed layout issues on larger screens for the public conversation page.

2. Chatbot Shrinking Fix:

   • Resolved the issue of the website chatbot shrinking on certain screens.

3. Embedding Option Typo Fix:

   • Corrected a typo in the embedding option.

4. Dashboard Styling Fixes:

   • Fixed small access to styling on the agency dashboard.

5. Resubscription Fix:

   • Resolved an issue preventing customers from re-subscribing to the same plan.

6. Subaccounts List Popover Fix:

   • Fixed the issue with the popover in the Sub-accounts list.

1. Free Trial Warning Fix:

   • Resolved a bug related to the free trial warning message.

2. Custom Function History:

   • Maintain a log of all custom function calls, allowing users to review past actions.

3. Email Address Change via OTP:

   • Agency users can now change their email addresses using OTP verification.

   

4. Password Reset Notifications:

   • Both sub-accounts and agencies receive alerts when a sub-account resets their password through the Stammer app.

5. Support Link Customization:

   • Agencies can personalize the support link displayed in the top header.

   

6. Enhanced Conversation Reset:

   • Improved reset feature for agent widgets to boost user experience.

7. Display Last Scraped Date:

   • Shows the last scraped date for each URL within the system.

   

8. Voice Functionality Update:

   • Voice feature is disabled in widgets when using ElevenLabs text-to-speech.

   • Google Text-to-Speech (gTTS) is used as a fallback if ElevenLabs is unavailable.

9. Feature Visibility Control:

   • Disabled features are only shown if provided by agencies; otherwise, they are hidden.

10. Disable Billing for Sub-Accounts:

    • Sub-accounts can be directly billed while using agency resources, including agents, wallet balance, and storage.

      

11. WhatsApp Integration Improvement:

    • Currently supports only text messages (audio coming soon)

New option to “Extract only main content with AI” will allow you to turn on/and off the ability to scrape all the text or just the main text from a webpage.

All Content = All content on the webpage which will result in a higher amount of characters scraped.

Main Content = Excludes duplicate data like menu items and repeat links which will result in a lower amount of characters scraped.

Agency Admin Permissions

You can now give additional admin permissions to another user inside your agency account. This will give this user full access to everything as the account owner except being able to cancel an active subscription or being able to delete the account owner.

🚀 1. ElevenLabs Integration for Full SaaS

• Full SaaS users can now input their own ElevenLabs API key for voice output, allowing AI Agents to generate custom voice responses.

  

📚 2. Rename Update page to settings

• Our old ‘Update’ section is now named ‘Settings’.

  

💬 3. AI Agent Daily Analytics Summary

• Introducing a new option with conversation analytics, detailed stats, and analysis, with results sent via email or webhook.

  

🎨 UI Improvements:

• Removed the logo in onboarding flow.

• Fixed a visual error a chatbot button.

🎯Other fixes:

• Custom function with same alias.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

Note: Features labeled as "in active development" may have bugs or undergo changes. Your feedback is crucial for improvement.

Human Handoff for Integrations

Upon detecting specific keywords, the AI agent stops responding and alerts support staff via email or webhook, ensuring prompt human assistance for users.

Create Unlimited Packages

Agencies can now create unlimited packages to sell to their clients. This also gives the ability to show/ hide packages to your clients.

Other Updates

• Users can scroll up while the AI agent is responding

• When a user clicks the "X" on a starter message, it will just close out the starter message and NOT open the bot.

🌐 1. Facebook/WhatsApp Integration:

• Limited to one Facebook page per agent

• Added explanations for required permissions

  

📚 2. Knowledge Base Enhancement:

• URLs of source pages now included in match results

• Improves ChatGPT's ability to reference information sources

💬 3. WhatsApp Integration:

• Now live (in active development)

• Beta testers needed

  

🛠️ 4. Custom Functions Prototype:

• Available in Labs area

• Allows bots to call URLs and receive responses

• In active development; interface & functionality may change

  

🔗5. Public Link for Conversations:

• New feature to share conversations via public links

  

🎨 UI Improvements:

• Truncated account names for better UI display

• Fixed voice functionality icon display issue

• And a lot more!

🎯Ongoing Development:

• Facebook integration

• WhatsApp integration

🚀 Coming Soon:

• New interface for chatbot communication

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

Note: Features labeled as "in active development" may have bugs or undergo changes. Your feedback is crucial for improvement.

🌐 1. Facebook Messenger Integration

The Stammer.ai uses the Official Facebook integration with Meta's API. You can connect any AI to any of your connect Facebook Pages. Simply sign in and select the page you want to use. A Facebook page can only have one AI agent connected at a time.

Step 1: Select an AI Agent from your dashboard

Step 2: Navigate to the Integrations tab

Step 3: Click on the Facebook integration

Step 4: Create/ Connect your Facebook Business account

Step 5: Enable the AI Agent to be active

You're done! The integration is now setup and when you message the Facebook Messenger, in a few seconds the AI agent will respond.

💬 2. WhatsApp Integration

The Stammer.ai uses the Official WhatsApp integration with Meta's API. You sign into your Facebook account and then will be prompted to either create or connect an existing WhatsApp number. Business WhatsApp phone numbers from Meta may have an additional expense depending on your location and sending category billed directly from Meta.

Step 1: Select an AI Agent from your dashboard

Step 2: Navigate to the Integrations tab

Step 3: Click on the WhatsApp integration

Step 4: Create/ Connect your WhatsApp Business account

Step 5: Enable the AI Agent to be active

You're done! The integration is now setup and when you message the WhatsApp phone number, in a few seconds the AI agent will respond.

🛠️ 3. Custom Functions

• Available for Full SaaS Mode users only.

• This is advanced functionality we are sharing with full SaaS mode users early to help test and improve UI/ UX.

• Custom Functions allow developers to create custom actions and API calls for their AI agents similar to creating a custom GPT.

  For example: “I want to create an AI agent that is connected to my Shopify inventory and provide live availability.”

• You could create a custom function which connects your Shopify store and gets all the data as needed for your use case so the AI agent can respond as needed.

🐛 Bug Fixes:

• Several other bug fixes and smaller platform updates were added as well.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

🎤 1. Voice Functionality:

• Location: Chatbot settings > Update tab > Advanced settings (at the bottom)

• Features: a) Microphone button in chat widget for voice input b) Text-to-speech output for chatbot responses

• Uses free web API built into most browsers (except Firefox)

• Video tutorial coming soon from Ross

  

🤖 2. AI Agency Internal Support Bot:

• Access: AI Agent tab > Internal AI Support Bot

• Only visible in white-labeled platform version

• Enhance internal support capabilities

  

📅 3. Conversational Scheduling Improvements:

• Fixed: Full-day busy events in Google Calendar now properly reflected

• Action Item: Test using getAvailability function in UI

• Report any discrepancies or issues via Discord: https://stammer.ai/discord

🌐 4. Meta Integration Update (in progress):

• Fixed: Facebook pages enabled via Meta Business Suite now appear in drop-down

• Facilitates resubmission to Meta for application approval

📝 5. Updated Default Base System Prompt:

• Applies to newly created agents

• Review and provide feedback on the new default prompt

🚀 Coming Soon:

• GPT 4o Mini model:

  • Crazy cheap!

  • More cost-effective than current GPT 3.5 model 🤯

• Release expected later today (July 18, 2024)

• Action Item: Test and provide feedback via Discord: https://stammer.ai/discord

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

Definition: Storage (characters) refers to the amount of text uploaded to an AI agent's knowledge base. Characters are the unit of storage. This definition is 154 characters long.

New Pricing Plans

Only new subscribers will be affected by the pricing changes. Existing subscribers can keep their current plan or switch to a new tier at any time.

🌟 Starter Plan ($49/mo) 🌟

The starter plan is now the most basic version of Stammer.ai without white labeling, built for agencies who need more time to build/ test their AI agents before selling.

• ✅ 5 AI Agents

• ✅ 5M Characters (Storage)

• ✅ Custom Knowledge Base

• ✅ Website Chat Widget

• ✅ All Integrations

• ✅ Unlimited Users & Contacts

• ✅ $25 Credit at Sign Up (used for wallet starting balance)

• ❌ White Labeling Not Available in Starter Plan

🚀 Agency Plan ($197/mo) 🚀

The agency plan is now the most affordable version of Stammer.ai with ALL white labeling functionality, built for agencies just getting started or those with a smaller client list.

• ✅ Everything in Starter

• ✅ All White Labeling Features

• ✅ 20 AI Agents

• ✅ 20 Client Sub-Accounts

• ✅ 20M Characters (Storage)

• ✅ Custom Functions

• ✅ API & Webhooks

• ✅ $100 Credit at Sign Up (used for wallet starting balance)

• ❌ Bring Your Own LLM API Key (OpenAI)

💎 Full SaaS Mode ($497/mo) 💎

The Full SaaS Mode plan is easily the most powerful version of Stammer.ai with the ability to sell unlimited AI agents to unlimited clients while using your own OpenAI API key to increase your profitability even further.

• ✅ Everything in Agency

• ✅ Unlimited AI Agents

• ✅ Unlimited Client Sub-Accounts

• ✅ 50M Characters (Storage)

• ✅ Bring Your Own OpenAI API Key 🚀

• ✅ $250 Credit at Sign Up (used for wallet starting balance)

• ✅ Access to New Features

Cost per AI Message - Updated (~40% Cheaper)

We also updated and simplified the cost of AI messages and other add-ons available for you to resell to your clients at a higher profit margin.

📩 Extra Messages:

• GPT-3.5: $0.005/message.

• GPT-4 Turbo: $0.05/message.

• GPT-4o: $0.02/message.

🤖 Extra AI Agents:

• $10/month per agent (note: Full SaaS Mode users have unlimited AI agents included at no additional cost).

💾 Extra Storage:

• $5/month (taken from wallet balance) per 1M characters.

💵 Wallet System Explained

The wallet system replaces the old message credits system. Old message credits will be converted to an available wallet balance at $0.003 per message credit for both agencies and sub-accounts. (Example: 500K message credits becomes a wallet balance of $1,500)

💰 Agency Wallet System 💰

The agency wallet balance is maintained by the agency's credit card with both manual and auto billing options.

• Example: When your wallet balance is less than $10, your credit card will be charged $10.

When is the Agency's Wallet charged?

The agency's wallet is charged for AI messages sent, extra storage and extra AI agents. A small amount for each AI message sent (or you can upgrade to the new Full SaaS Mode to use your own OpenAI API key and avoid this cost).

Cost per AI Message Sent

• GPT-3.5: $0.005/message.

• GPT-4 Turbo: $0.05/message.

• GPT-4o: $0.02/message.

Cost per Extra Storage (Characters)

• $5/month for 1,000,000 characters.

Cost per Extra AI Agent (Not applicable for Full SaaS Mode since they have unlimited AI agents already)

• $10/month for 1 AI Agent.

When is the Agency's Credit Card charged?

• When you increase the balance of your wallet manually.

• When you set your auto-billing threshold and your wallet balance falls below that threshold.

Why is Stammer making this change to a wallet system?

• We work extremely closely with our subscribers and community (on a daily/ weekly basis with our Discord and Office Hours) to figure out how to improve Stammer.ai. One of the biggest pieces of feedback we've gotten in the last few months is how confusing message credits are to explain to customers. In order to solve this problem we had to redesign how the platform works. Now, with a wallet system we can provide a better user experience and have a better system for usage billing based features like voice, image generation etc. The wallet is just the beginning of the new features that we built based directly on feedback from our customers who helped us refine it along the way.

🏦 Client Sub-Account Wallet System 🏦

• Each client has a sub-account wallet maintained by the client’s credit card.

• This balance is used to pay for each AI message sent at the price your agency has set.

• Example: When your client's AI agent (using GPT-4o) sends a message, your agency wallet will be charged $0.02 and at the same time the sub-account's wallet will be charged $0.05 (or whatever price you set as the profit margin for GPT-4o messages).

When is the Client Sub-Account's Wallet charged?

The client's wallet is charged for AI messages sent, extra storage and extra AI agents inside their sub-account based on the prices set by the agency.

When is Client Sub-Account Credit Card charged?

• When the client increases the balance of their wallet manually.

• When you or your clients set the auto-billing threshold for the sub-account wallet and wallet balance falls below that threshold.

These payments will charge the client's credit card and are sent directly to the agency's Stripe account as new revenue/ income.

🏗️ Updates to the SaaS Configurator

What is the SaaS Configurator?

• The SaaS configurator allows agencies to build custom subscription packages to use to organize and build offers for clients. After taking feedback from our customers we have updated it substantially to include a ton of new features and control over the different pricing options for each package.

The new SaaS configurator inside Stammer.ai allows agencies to have advanced control over the creation and management of subscription packages and what AI features to include to sell to clients.

These subscriptions will have all of these different customization options.

Plan Details Tab:

• Package Name

• Package Tier

• Package Description

• Package Features

• Number of AI Agents

• Number of Characters for Storage

Pricing Tab:

Agencies have the ability to create multiple different price options for each of their packages. There will be an enable/disable option for each type of subscription so you control what your clients will see.

• Monthly Price

• Quarterly Price

• Yearly Price

Features Tab:

Agencies now have the ability to set what permissions they want to give (or not give) to any client who subscribes to that specific package tier.

• Permission Options:

  • Create Chatbot

  • View Chatbot Dashboard

  • Access Chatbot Settings

  • Delete Chatbot

  • Access Knowledge Base

  • Access Leads

  • Access Scheduling

  • View Conversations

  • Customize Appearance

  • Access Integrations

  • Access API

Trials & Credits Tab:

Activate/Deactivate Free Trial:

• You can set a free trial period for your subscription plans and choose how many days the trial lasts. When your clients sign up, they will enter their credit card details but won't be charged until the trial period ends.

Complimentary Credit:

• You can give free credit when clients purchase a subscription plan, either as a one-time credit or recurring monthly credit. This credit will be added to the client's sub-account wallet balance.

Message Re-billing Tab:

This page shows how to set markup margins for different AI message models, allowing you to control the cost charged to customers and the profit per message for each model.

🤝 For Existing Subscribers

First of all, thank you for your continued support!

We've heard your feedback so here's all the goodies we made for ya:

1. Your subscription price will stay the same or get cheaper (depending what you're paying right now.

2. Everything is going to work the exact same you just now have more control. All of your AI agents (and your sub-account's AI agents) will continue to work the exact same way.

3. Message credits will be being converted into a wallet balance at the rate of 1 Message Credit = $0.003. You will have a Agency Wallet Balance. Your Clients will have Sub-Account Wallet Balance.

4. Usage is getting cheaper. Instead of "5 message credits" for a GPT-4o message, we are simplifying to a 'cost per message' that will be deducted from your agency wallet balance. GPT-4o = $0.02 /message. GPT-4 Turbo = $0.05 /message. GPT-3.5 Turbo = $0.005 /message.

5. AI agents inside each client sub-account will have a balance that is used to pay for each AI message sent. As an agency you will set the profit margin for each AI message sent for each of your clients.

6. You will have a Storage Balance of characters. This is the amount of data that you and your clients can upload to your AI agent's knowledge base. You can upsell additional storage to your clients and keep 100% of the profits.

7. With this update you have more control and customization over the subscriptions you sell to your clients.

8. If you signed up before 7/10/24 you will receive a bonus of up to $1,000 added to your wallet balance and a bonus of 1,000,000 characters added to your storage balance. Schedule meeting to get your bonus.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

Speak with the Stammer.ai Team

https://calendly.com/stammer-ai/stammer-1-5-review

📄 1. Sub-account display:

• Dashboard now shows sub-account name instead of "sub-account admin" label

  

📅 2. Calendar integration:

• Connected Google account email now visible in scheduling tab

  

🔑 3. API key management:

• Added confirmation step before deleting API keys

✅ 4. QA management:

• New batch deletion option for QAs

• Can now delete all QAs at once

🎨 UI Improvements:

🤖 1. Main chatbot list view:

• Updated display

• New sorting options

• Removed inner scroll bars

💬 2. Chat widget:

• Improved local storage system functionality

🎯 Additional Updates:

• Various minor bug fixes and backend improvements

🚀 Upcoming:

• Stammer v1.5.0 scheduled for release on Wednesday, July 10th

• Planned knowledge base downtime: July 10th, 5-8 AM PST (Chatbots will remain functional during this period but will not be able to access the knowledge base)

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

📄 1. Markdown files

Added support for uploading and processing markdown files.

🗂️ 2. New sorting options for chatbot list view:

1. Sort by name

2. Sort by creation date

✅ 3. QAs in batch

• Added ability to select and delete QAs in batch

🎨 4. Background color of chat widget

• New option to change background color of chat widget input area

🐛 Bug Fixes:

• Corrected issues in test lead webhook functionality:

  • Successful testing now possible

  • Test format now matches actual webhook data format

🎯 Additional Updates:

• Various minor bug fixes and improvements.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

🔑 Key Features:

🆕 1. GPT-4o as default mode

• GPT-4o now set as default model for new chatbots.

🐛 Bug Fixes:

• Resolved issue with Google indexing protected pages.

• Corrected data format display in webhook history/test feature.

🎨UI Changes:

• Renamed "Send Test Data" to "View History and Test" under Leads tab.

🎯 Additional Updates:

• Various minor bug fixes and improvements.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

🔑 Key Features:

🔧 1. Chat Widget Auto-Opening:

• Resolved an issue where the chat widget would open multiple times on a website.

🆕 2. Conversation View Updates:

• Updated conversation view IDs to be sequential for each chat bot's conversation number, improving clarity.

• Added the platform origin display in the conversation view tab, indicating whether the message came from WhatsApp, the dashboard, or another source.

📧 3. Email Updates:

• Updated terminology and formatting in several emails based on customer feedback.

• Resolved a minor issue causing the duplication of certain emails in an edge case.

🔄 4. Base System Prompt Reset:

• Introduced the ability to reset the base system prompt back to default.

📃 5. Subaccounts List View:

• Improved and updated the user interface of the subaccounts list view page for better usability.

We are working on significant new features and functionality, so stay tuned for exciting updates coming soon.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

🔑 Key Features:

🔧 1. OpenAI Status Message Fix:

• Fixed persistent OpenAI status message in the agency dashboard.

📧 2. Email Duplication Issue:

• Resolved rare duplicate email occurrences.

🚀 3. Dashboard Improvements:

• Updated text, terminology, and language.

• Made several formatting and styling adjustments.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

🚀 1. GPT-4 Turbo model updated (GPT-4-Turbo-2024-04-09)

• Experience the next generation of AI with our latest model update.

• Automatically upgraded for all existing GPT-4 Turbo users to ensure top-notch service.

  

🆕 2. NEW! GPT-4o model

• Explore our new model option and discover its unique capabilities.

• Only 5 credits per response

🐛 Bug Fixes:

Resolved an issue that affected some lead gen bots earlier in the week

🎯 Interface & Additional Updates:

Various minor improvements and updates to the user interface.

Back-end improvements in preparation for upcoming features.

💡 Feedback & Support:

The team at Stammer is excited about the community involvement and looks forward to continued feedback and collaboration on future features and improvements.

Join the Discord

https://stammer.ai/discord

Join Next Office Hours (Tuesdays 11am-12pm pst)

https://stammer.ai/officehours

🕒 1. 24-hour scraper

• 24-hour scraper now available on all plans, including free trial.

🎨 2. Dashboard customization

• Dashboard customization with five selectable themes.

✅ 3. API endpoint

• Added API endpoint for adding a QA item.

💬 4. Conversation Starters

• Enabled persistent conversation starters (suggested messages).

🗣️ 5. Prompt variables

• Introduced prompt variables, starting with current time in UTC.

🐛 Bug Fixes:

• Resolved minor issues with dark mode.

• Fixed an issue with text being too small on a certain user's web page.

• Ongoing user interface enhancements, such as progress bars for chat bot usage.