Skip to content

v2.9.0

Latest
Latest

Choose a tag to compare

View all tags
@rnabeelTT rnabeelTT released this 28 Jul 18:09
· 4 commits to main since this release
rc-v2.9.0
8d622a0
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

🚀 New Features

  • OpenClaw Integration: Added OpenClaw as a deployable general agent alongside the coding agents lineup
  • Workflow Canvas Upgrades: Model auto-select, save/new flow, and live code streaming, plus dynamic edge styling, edge hover-delete with guards on required nodes, loaded-template name in the toolbar, and auto-selection of the first RAG collection
  • In-App First-Run Setup: Moved first-run setup and runtime secrets into the UI, so keys and config no longer require touching .env
  • Editable Voice Agent Prompt: Voice Agent system prompt is now editable at runtime
  • RAG Relevance Controls: Added query-relevance controls for RAG ingestion and retrieval
  • Grouped Navbar: Desktop navbar items organized into Models, Tools, and Playground dropdowns
  • Launcher --uninstall & --switch: New teardown/branch-switch commands plus stale tt-studio shortcut auto-repair
  • Backend-Unreachable Overlay: Blocking overlay shown when the backend is unreachable

🤖 Model Support

  • Flux Schnell: Added Flux Schnell to the model catalog
  • Health-Gated Navbar Entries: Model navbar entries appear only once the model is healthy
  • Parallel Deployments: Model deployments now run in parallel before the container-start phase

🐛 Bug Fixes

  • Fixed external model detection and registration
  • Fixed JWT secret detection and tool-calling for registered models
  • Fixed RAG ingestion bugs
  • Welcome/first-run setup now persists across restarts
  • get_env_var now prioritizes os.environ over .env
  • Speed Insights now shows each chat message's own model
  • Replaced a hardcoded FastAPI gateway IP
  • Voice Agent components now labeled by function instead of model name

🔧 Technical Improvements

  • Bumped tt-inference-server artifact to v0.18.0
  • Refactored run.py into the tt_setup package and consolidated launcher config dotfiles into a single store
  • Consolidated duplicated chipStatus types into a shared interface
  • Kept backend runtime state out of the bind-mounted source tree
  • Auto-install frontend npm deps on dev-container start; removed the sudo password prompt from setup
  • Added CI: frontend pipeline, launcher test suite, and deploy healthcheck
  • --purge-all now removes the launcher venv

Full changelog → #1157

Assets 2
Loading