Skip to content

v2.0.0

Choose a tag to compare

View all tags
@github-actions github-actions released this 21 Jul 01:45
· 39 commits to main since this release
v2.0.0
50fa676

2.0.0 (2026-07-21)

⚠ BREAKING CHANGES

  • llm: the intent_judge_model, tool_router_model,
    evaluator_model, and planner_model config keys are removed
    (auto-migrated to fast_model where explicitly set).

✨ Features

  • frictionless multi-provider setup (discovery, capability probe, startup check) (#482) (1d5ab5d)
  • llm: add OpenAICompatibleBackend, llm_provider config, factory dispatch (#389) (e2d4a72)
  • llm: two-tier model system (fast model + chat model) (#525) (70a7fed)
  • ui: select OpenAI-compatible LLM provider from the UI (#444) (a5fae58)
  • wizard: guided model selection for OpenAI-compatible setup (#457) (7dd71ce)

🐛 Bug Fixes

  • audio: serialise PortAudio stream lifecycle across threads (#521) (de1499c), closes #462 #401 #422
  • config: expand tilde in user-supplied path settings (#520) (2adaec4), closes #467
  • config: make config.json writes atomic (#529) (d90af34)
  • config: rename llm_embed_timeout_sec to llm_embedding_timeout_sec (#390) (330dcfa)
  • desktop: keep wizard worker threads alive until fully finished (#517) (20d1795), closes #509 #407 #239
  • dictation: survive missing PortAudio/pynput at import time (#515) (65f4fd2), closes #503
  • harden warmup probes, warm embed model, and show wizard on auto-start failure (#530) (294b231)
  • llm: provider-aware auxiliary model resolution (#516) (2832729), closes #363
  • strip non-standard message fields before sending to LLM API (#532) (50fa676)
  • wizard: provider choice first, exclusive radios, clearer copy (#446) (e57f133)

📝 Documentation

  • fix stale Ollama-only prerequisite messaging (README + release notes) (#528) (ed476fd), closes #457 #525
  • record offline-first principle (no vendor-locked cloud integrations) (#480) (33065ad), closes #429

♻️ Code Refactoring

  • llm: extract pluggable LLMBackend abstraction (#388) (47e4d4b)
  • llm: route every call site through factory dispatch (#392) (3b9bfd5)

⚡ Prerequisites

  • Ollama (default), or any OpenAI-compatible server (LM Studio, Jan, llama.cpp, vLLM, oMLX, LocalAI, …)

📦 Downloads

Platform File Notes
Windows Jarvis-Windows-x64.zip Extract → Run Jarvis.exe
macOS (Apple Silicon) Jarvis-macOS-arm64.zip Extract → Move to Applications → Right-click → Open
macOS (Intel) Jarvis-macOS-x64.zip Extract → Move to Applications → Right-click → Open
Linux Jarvis-Linux-x64.tar.gz tar -xzf → Run ./Jarvis/Jarvis
Assets 6
Loading