Changelog

June 2026

Transparency & Source Insights

• Improved source visibility by allowing administrators to expand source chunks and view the complete content used to generate AI responses.

• Continued improvements to response traceability and answer validation workflows.

Training & Content Management

• Improved source training and retraining reliability, including better recovery from failed training runs and storage quota interruptions.

• Improved handling of quota-related training scenarios to ensure source processing behaves more consistently.

Chat & User Experience

• Updated default link behavior so chatbot links open in the current tab unless configured otherwise through advanced deployment settings.

• Improved consistency of feedback collection across administrative and testing interfaces.

Stability & Reliability

• Fixed several issues affecting training workflows, feedback submission, and administrative operations.

• Continued improvements to platform reliability, validation, and error handling.

May 2026

Agent Management

• Added Agent Avatars!

• Added ability to enable/disable agents on the fly

Onboarding & Signup Improvements

• Improved the onboarding experience with streamlined agent and source creation during setup.

• Enhanced the mobile signup experience for better usability on phones and tablets.

Chat & User Experience

• Improved analytics compatibility by allowing chatbot events to propagate correctly to Google Analytics and Tag Manager integrations.

• Fixed an issue where starting a new chat could interrupt the response of an active conversation.

• Fixed issues where website styles could leak into the CrafterQ UI and affect chatbot appearance.

• Fixed an issue where explainer videos would fail to play on iPhone devices using Chrome.

• Improved handling of websocket events to keep the UI more accurately synchronized with system activity.

• Fixed issues where dashboard views could display an extra day in reporting.

Playground Improvements

• Added source information visibility in Settings View (soon in Playground) so admins can better understand how answers are generated.

Billing & Subscription Improvements

• Fixed subscription upgrade and add-on purchase flows for payment methods requiring 3D Secure (3DS) authentication.

• Improved subscription lifecycle handling for organizations with overdue or failed payments.

AI & Response Quality

• Improved prompts so organizations can optionally allow the AI to use broader model knowledge beyond trained sources when configured to do so.

Community & Identity

• Improved Community usernames so they are more stable, identifiable, and user-friendly.

Platform Reliability

• Improved signup notification infrastructure and related environment configuration.

April 2026

Billing & Subscription Improvements

• Improved accuracy and consistency of subscription information, including:

  • clearer display of current plan (monthly vs annual)

  • improved next billing date handling

  • better synchronization between the UI and payment provider

• Improved behavior when updating or changing payment methods.

• Fixed issues where subscription changes required manual page refresh.

• Improved handling of subscription lifecycle events, including renewals and plan transitions.

• Added safeguards to prevent unintended plan downgrades during checkout or billing updates.

• Improved visibility and handling of cases where no valid payment method is present.

Admin Experience Improvements

• Enhanced account management experience, including improved navigation and form behavior in the My Account page.

• Improved focus handling and usability across account-related screens.

• Improved prompt editing experience for agents, making it easier to configure behavior.

• Ongoing updates to the Playground UI for a more consistent and usable testing experience.

• Improved error messaging (e.g., clearer feedback when creating users that already exist).

Chat & Agent Behavior

• Improvements to prompt handling to make agent responses more consistent and predictable.

March 2026

Admin Experience Improvements

• Added more control over chatbot presentation, including Chat UI configuration and improvements to UI modes.

• Source settings are more flexible, including the ability to control whether refresh is enabled per source.

• Added a My Account page and improved account-related flows, including clearer password policy messaging and better reset-password feedback.

• Improved dashboards and chat logs with fixes to filtering, date ranges, and graph display issues.

Chat Experience Improvements

• Improved chatbot reliability so end-users are less likely to lose responses during intermittent inference issues.

• Added a minimize button so end-users can collapse the chatbot back to its compact launcher.

• Improved follow-up question visibility in dark mode.

• Improved conversation handling so Q&A content is better considered when detecting topic changes.

Security & Access

• Tightened role-change permissions to better protect administrative access.

• Improved allowed-domain validation to support more realistic configurations, including wildcards and multiple ports.

• Strengthened security-related configuration around allowed origins.

Training & Source Management

• Improved source training reliability, including fixes for:

  • include/exclude link patterns

  • retraining triggers

  • large Q&A source updates

  • crawling limits

  • problematic crawling scenarios that could miss links or loop incorrectly

• Improved visibility of training progress by handling new training start events in the UI.

• Fixed cases where source dialogs or upload retry flows could get stuck or behave confusingly.

February 2026

Security

• Allowed domains is now a configuration item per agent to limit where the agent is allowed to run.

• New login experience.

Stability & Reliability Improvements

• Improved overall system reliability during training, retraining, and chat usage.

• Fixed multiple issues where retraining or crawling could stall, fail silently, or not trigger as expected.

• Better handling of temporary connectivity issues in the Chat UI, including automatic retries.

• Improved session handling so users don’t lose their place when sessions expire.

Chat Experience Improvements

• Better conversation flow, reducing irrelevant or repetitive answers.

• Follow-up questions are now more accurate and no longer appear when there is no meaningful answer.

• Reduced cases where the assistant fixates on past topics or repeats prior responses.

• Improved handling of short responses like “ok” so they don’t trigger repeated answers.

• Responsive chat client that turns into a little bubble when the site is viewed on mobile devices.

• Configurable chat client so you can decide the mode and position (conversational vs bubble).

Training & Content Improvements

• Training is now more reliable when crawling sites with redirects, sparse sitemaps, or large link structures.

• Improved crawling logic to avoid infinite loops and missed pages.

• Support for training on linked documents (such as PDFs and text files) found during crawling.

• More accurate tracking of when sources were last updated.

• Retraining can now be triggered directly from the Sources menu.

• Configurable auto-refresh so you can turn automatic refresh on/off per web source.

UI & Usability Enhancements

• Cleaner, simpler agent creation and source setup flows.

• Improved spacing, layout, and readability across dialogs.

• First-time user experience improved with clearer guidance and actions.

• Better scroll behavior and focus handling, including fixes for Linux and mobile browsers.

• Password rules are now clearly shown in the UI.

• The app no longer jumps or resets views unexpectedly during common actions.

Bug Fixes

• Fixed issues where buttons (Create, Refresh, Retrain) appeared to do nothing.

• Resolved cases where uploads failed due to long filenames or unsupported formats.

• Fixed issues where changes didn’t immediately reflect in dashboards, lists, or logs.

• Improved handling of expired sessions to prevent confusing errors.

January 2026

Private Beta Launch

• CrafterQ entered private beta.

• Core functionality for agents, training, crawling, and chat was stabilized.

• Initial onboarding, documentation, and changelog access added to the app.