Latest Release Highlights
Current feature guides with previous release highlights kept for reference
This page mirrors the curated in-app Latest Features experience and keeps earlier release highlights available in an archive section below the current release.
Current Release Features
These guides map to the feature set currently highlighted in the app for end users.
Guided Tutorials
Step-by-step walkthroughs help users discover core chat, workspace, and onboarding flows in context instead of hunting through the interface first.
The fastest way to learn a new workflow is usually inside the workflow itself, with the right controls highlighted as you go.
Background Chat
Long-running chat requests can finish after you leave the chat page, so larger prompts and uploads stop blocking the rest of your work.
Long-running work no longer makes the app feel blocked while the assistant is still processing.
GPT Selection
Visible model selection helps users match a task to the right tradeoff in speed, cost, and reasoning depth when multiple options are available.
Different prompts often need different tradeoffs, so a clearer picker makes model choice practical instead of hidden.
Tabular Analysis
CSV and spreadsheet workflows keep getting stronger for grounded exploration, filtering, and follow-up analysis against real rows and columns.
Table-heavy questions are more useful when users can explore filters and follow-ups without losing provenance.
Citation Improvements
Enhanced citations improve source traceability, document previews, and history-aware grounding so users can inspect where answers came from.
Traceable answers are easier to trust, revisit, and defend in follow-up conversations.
Document Versioning
Same-name uploads now behave like revisions, making it easier to keep document history visible instead of silently replacing useful versions.
Revision-aware documents keep history available when teams need to compare or verify what changed.
Summaries and Export
Conversation summaries and export actions keep expanding so users can package, review, and share useful chat outcomes more quickly.
People reuse chat outcomes faster when they can summarize and package them for someone else without manual cleanup.
Agent Operations
Agent creation, organization, and operational controls keep getting smoother for teams that rely on more advanced agent and action scenarios.
Advanced agent setups are easier to adopt when browsing, defaults, and validation are clearer.
AI Transparency
Thought and reasoning transparency options give users clearer feedback about what the assistant is doing while a response is being prepared.
Visible processing cues reduce uncertainty while the system is working and make delays easier to interpret.
Fact Memory
Profile-based memory now separates durable Instructions from recall-only Facts so the assistant can carry preferences and relevant context more cleanly.
Separating durable preferences from recall-only facts keeps memory more relevant and less noisy over time.
Deployment
Deployment guidance and diagnostics continue to improve so admins can roll out changes with less guesswork and better operational feedback.
Rollouts are safer when guidance, settings, and diagnostics are easier to follow under pressure.
Redis and Key Vault
Caching and secret-management guidance has expanded so production deployments can be configured more securely and more predictably.
Secure caching and secret handling are easier to operationalize when the guidance is concrete instead of implicit.
Send Feedback
End users can prepare structured bug reports and feature requests directly from the Support menu instead of sending unstructured notes.
Structured feedback gives admins something actionable instead of vague complaints or one-line requests.
Support Menu
Admins can expose Latest Features and Send Feedback through one predictable support surface so discovery and reporting live in the same place.
A predictable support surface makes help, announcements, and feedback easier to discover and reuse.
Archive Previous Release Features The earlier v0.239.001 release guides remain available for reference. v0.239.001 Show highlights
Export Conversation
Users can export one or multiple conversations in JSON or Markdown through a guided wizard that handles packaging and download choices.
Structured export makes it easier to archive, review, reuse, or share conversations outside the live chat UI.
Retention Policy
Retention settings let admins and workspace owners control how long conversations and documents are kept before automatic cleanup runs.
Retention controls make lifecycle management more predictable, auditable, and easier to explain across workspaces.
Workspace Scope Lock
Workspace scope locking freezes the originating workspace selection after grounded search so follow-up questions do not accidentally cross data boundaries.
Locking the originating scope prevents accidental cross-contamination when users continue a grounded conversation.
Tags, Grid View, and Chat Filtering
Tag management, folder-style grid view, and chat filtering work together so users can organize large document sets and search them more intentionally.
Tags and folder-style browsing turn large document collections into something people can navigate and filter quickly.
Additional highlights from v0.239.001
- Owner-only group agent management lets admins restrict group agent and action edits to group owners.
- Multi-workspace scope management expanded grounded chat across personal, group, and public sources.
- Tag folders, grid view, and chat filtering shipped together as one broader workspace workflow upgrade.
Bug fixes kept for reference
- Citation parsing fixes resolved edge cases where page range references failed to create the right clickable links.
- Public workspace activation fixed a 403 path for non-owner users activating a public workspace for chat.