- Jan 14, 2026
- Parsed from source:Jan 14, 2026
- Detected by Releasebot:Jan 15, 2026
2026-01-14
Supported Agent Skills in SOLO mode
- Supported uploading and adding Skills to extend agent capabilities
- Enabled creating Skills directly within conversation flows
Supported Memory functionality in beta
- Supported global and project-level memory types
- Enabled toggling memory functionality on/off
- Allowed manually adding or removing memory entries
Added DevTools direct access in Browser preview
Improved Mermaid diagram rendering in SOLO mode DocView
Original source Report a problem - Dec 29, 2025
- Parsed from source:Dec 29, 2025
- Detected by Releasebot:Dec 31, 2025
- Dec 26, 2025
- Parsed from source:Dec 26, 2025
- Detected by Releasebot:Dec 31, 2025
2025-12-26
- Cue upgraded to Cue-Pro, added Repository-Level Intent Understanding and Cue-Pro in the sidebar to make edit predictions visual and navigable.
- Dec 25, 2025
- Parsed from source:Dec 25, 2025
- Detected by Releasebot:Dec 26, 2025
2025-12-25
Supported jumping to a specific file or directory from the chat flow.
Supported opening the Supabase tool in an individual tab in SOLO mode.
Added Vercel AI Gateway as an AI service provider for SOLO mode.
- Dec 22, 2025
- Parsed from source:Dec 22, 2025
- Detected by Releasebot:Dec 23, 2025
2025-12-22
TRAe upgrades command execution workflow with three configurable modes: Sandbox with Allowlist, Manual Run, and Auto Run, now rolling out. Mermaid is supported in conversations, blending safety and speed for AI commands.
Command execution modes
To ensure secure command execution, TRAE has upgraded the workflow for running AI-generated commands. You can configure one of the following three command execution modes in the settings center (currently in gradual rollout):
Sandbox with Allowlist (enabled by default): Commands are executed automatically inside a secure sandbox with risk assessment. Commands included in the allowlist bypass the sandbox and are executed directly.
Manual Run: All commands require manual confirmation before they are executed.
Auto Run: All commands are executed automatically outside the sandbox.
Mermaid is supported in conversation.
- Dec 18, 2025
- Parsed from source:Dec 18, 2025
- Detected by Releasebot:Dec 17, 2025
- Modified by Releasebot:Jan 15, 2026
2025-12-18
Model Update
- Gemini-3-Flash-Preview is added as the built-in model in IDE mode.
- Dec 16, 2025
- Parsed from source:Dec 16, 2025
- Detected by Releasebot:Dec 17, 2025
- Modified by Releasebot:Jan 15, 2026
2025-12-16
Model Update
- Kimi-K2-0905 is added as the built-in model in SOLO.
- GPT 5.2 is added as the built-in model in IDE mode.
- Dec 12, 2025
- Parsed from source:Dec 12, 2025
- Detected by Releasebot:Dec 17, 2025
- Modified by Releasebot:Dec 23, 2025
2025-12-12
Feature Update
- Optimized the user experience of the AI input box.
- Optimized the user experience of the feature「Code Review」.
Model Update
- Improved the success rate of Remote-SSH connections.
- GPT 5.1 is added as the built-in model in IDE mode.
- Dec 10, 2025
- Parsed from source:Dec 10, 2025
- Detected by Releasebot:Dec 23, 2025
- Modified by Releasebot:Jan 15, 2026
2025-12-10
Model Update
- Gemini-3-Pro-Preview is added as the built-in model in SOLO.
- Dec 3, 2025
- Parsed from source:Dec 3, 2025
- Detected by Releasebot:Dec 17, 2025
v3.3.2
Release notes
- "Auto Accept" upgraded to "Code Review" settings both IDE and SOLO modes support 3 review options (Review all changes/Review latest changes/Skip review) for flexible AI code change control.
- Custom Model is supported in SOLO mode.
- Supported keyboard shortcuts to open Plan (Windows: Alt+P; macOS: Option+P) in SOLO Coder.