Skip to content

Version v.1.2.0 MCP!

Choose a tag to compare

View all tags
@GDKAYKY GDKAYKY released this 13 Mar 06:13
· 25 commits to main since this release
v.1.2.0
5eecabf
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Llama Desktop v1.2.0

image

What's New

image

MCP Integration

Connect any MCP server directly from the chat. Type / to open the command menu, pick a server, and the AI will automatically route your request through the right tool — no manual configuration needed.

Slash Command Menu

Typing / now opens a tool picker with quick access to MCP servers, skills, and commands.

Improvements

  • Streaming performance — markdown rendering is significantly faster during long responses
  • Thinking mode — Qwen3 chain-of-thought is now displayed inline with a collapsible block that auto-collapses once the answer arrives
  • Chat titles — conversation titles are now generated automatically from the first exchange
  • UI polish — refined chat layout, updated icons, and better visual consistency throughout

Fixes

  • MCP tools not loading after config save
  • <think> tags leaking into visible message content
  • Intent classifier failing to parse JSON when wrapped in think blocks
  • Chat title containing raw model output instead of a clean name

Infrastructure

  • Added SonarCloud analysis
  • Rust CI with full test coverage
  • Release pipeline for Windows NSIS installer
Assets 3
Loading