Welcome back to the AIsuru release notes for November 2025. Here you'll find new features, improvements, and bug fixes to make your AI Agents more powerful and secure ๐
NEW FEATURES AND IMPROVEMENTS!
- New languages - We have now successfully added 13 additional EU languages to our multilingual chat: Bulgarian, Croatian, Czech, Danish, Estonian, Latvian, Lithuanian, Dutch, Romanian, Slovak, Slovenian, Swedish, and Hungarian.
- Visibility and layout on mobile - Fixed visibility issues on mobile layout for many pages and table UIs.
- Under to "Tools" section in the "Create Function from file" it is now possible for our users to insert a JSON text.
- New layout settings - We added 8 new customizable widget attributes including copy button visibility, chat history display, auto-start functionality, and speaker settings.
- Code Snippet Display - Long code snippets (preformatted texts) now display as expandable media cards instead of taking up excessive chat space. Click to expand and view full content. With this way the user can have cleaner chat interface with better navigation and reduced scrolling for conversations containing large code blocks.
- Content Types for functions: we now support multipart/form-data in the webhooks of the advanced functions.
- We now support base64 images within the chat.
- IP Addresses in the conversations are now anonymized.
- From now on after the user enables a new MCP, the activated MCP will automatically be removed from the list with he available MCPโs
- It is also possible now to copy your own text in the conversation with your agent by simply going to the โshare meโ section selecting โ<\>โ ( embedded) and activating โShow copy buttonโ in the chat actions
- Our users can now only access the โ Unanswered questionsโ in the โContentsโ aerea only if they have the โGenerated AIโ disabled in โSettingsโ
- You can now create your very own SVG (Scalable Vector Graphics) by activating the โartifactsโ in the โtoolsโ section!
- From now own when a user creates a new agent they must also select the model for โConfiguration for question answeringโ as well as โConfiguration for import/exportโ
- Warnings now include MemoryID, enabling direct access to similar existing content.
- In the โchatโ aerea with your agent you wonโt be able to send a second text until the user receives a reply from the agent
- We have now a new improved date sector for for better User interface
- When a user creates a tool by uploading a document or a file, the generated URL now it is protected and not publicly accessible
- New option for Agents to prevent accumulating unanswered questions and AI-generated content
Bug fixes
- Fixed accidental message sending when hovering over the microphone icon. Messages now only send when intentionally submitted.
- Author Invitation - Author invitations with custom messages no longer trigger false "invalid email address" errors. Email validation now works correctly regardless of whether a personalized message is included.
- We now correctly apply tenant's CSS color scheme for the agents.
- Bulk delete functionality - Bulk delete now correctly targets only draft content, preventing accidental deletion of user-created content.
- Chat interface bug - Resolved overlapping elements in chat interface: the message input area and the file upload box no longer cover chat messages or each other.
- Bug fixes in both audio and visual system for the "TOTEM" layout.
- Model selection - Updated model selection labels in the Admin > Agents section to correctly show Board of Experts models details.
- Smaller improvements to Deep Thought - particularly to the research of interested users for a specific query in the Followers section.
- Admins of the tenant can now see the instructions of the Agents and can edit private Agents and Board of Experts (no invitation required).
- Fixed scrolling issues that were interfering with accessibility (mostly visible on mobile) in the areas โconversationโ and โsettingโ.
- Fixed an issue where in the โcontentsโ aera -> โAdd contentโ if the user selected the โsuggest answerโ from a piece of content, the reasoning model would also display the <think> tag.
- From now own the user can successfully upload a .txt file.
- Fixed overlapping UI elements from Manuela integration on the dashboard.
๐ ๏ธ As always, we've addressed various minor bugs and implemented performance improvements across the platform to ensure a smoother, faster experience.
๐ก Need help with AIsuru? Want to explore the AIsuru AI Academy? Contact us at demo@memori.ai