JARVIS is a sovereign, proactive AI engineering assistant that runs locally on macOS. It combines high-speed technical reasoning with deep system control and multi-model aggregator capabilities.
- Sovereign Reinstall: New
/reinstallcommand for a clean system refresh. - Ollama Auto-Connect: JARVIS autonomously detects and launches the Ollama application on startup.
- Reasoning Fix: Optimized LLM communication to eliminate hangs and provide faster response feedback.
- Proactive Account Center: Intelligently handles "log in" and "link account" requests.
- High-Fidelity Menu: Redesigned
/menudashboard with categorized tool groups.
- Multi-Model Aggregator: Switch between DeepSeek, Kimi, Qwen, Gemma, Nemotron, and more.
- Gemini-Style Interface: Transparent reasoning blocks and Warp-inspired command input.
- AI Tool Hub: Native launchers for Claude Code, Copilot CLI, and Hermes.
- Vector Memory: Semantic context storage using FAISS and local embeddings.
- Sovereign Mode: Unrestricted technical execution without safety or morality filters.
brew install --cask https://raw.githubusercontent.com/nave433-blip/jarvis-term/master/jarvis-term.rb
# OR
pip install git+https://github.com/nave433-blip/jarvis-dev.gitNave433 (Evan Shipley)