OpenJarvis

OpenJarvis

Personal AI, On Personal Devices

Build personal AI that runs on your hardware. Cloud APIs are optional.

Quick Start

# One command — installs everything for your hardware
$curl -fsSL open-jarvis.github.io/install.sh | bash

Runs entirely on your hardware — cloud APIs are optional. Browse and pull local models any time with jarvis model.Read the full docs →

Five primitives for personal AI

Intelligence

Pick a model, or let OpenJarvis pick one for your hardware. Manages the full catalog of local models across providers.

Engine

The inference runtime — Ollama, vLLM, SGLang, llama.cpp, and cloud APIs. Auto-detects your hardware and recommends the best fit.

Agents

Multi-step reasoning with tool use. Eight built-in agent types, from simple chat to orchestrated workflows.

Tools & Memory

Web search, calculator, file I/O, code interpreter, retrieval, persistent local state, and any external MCP server.

Learning

Your AI gets better over time. Every interaction generates traces that drive automatic improvements to weights, prompts, and agent behavior.

Runs on your stack

See all backends

OpenJarvis auto-detects your hardware and recommends the best-fit engine. The same code runs entirely on your machine or routes selectively to the cloud.

Engines
OllamavLLMSGLangllama.cppCloud APIs
Hardware
Apple SiliconNVIDIAAMDNPUsCPUs
Tools & protocols
MCP serversOpenAI-compatible APIWeb searchCode interpreterRetrieval

Join the community

OpenJarvis is open source and built in the open. Come say hi on Discord, star us on GitHub, or grab the desktop app and run your first local agent today.