# open-jarvis/OpenJarvis

Personal AI, On Personal Devices

Repository: https://github.com/open-jarvis/OpenJarvis
Canonical: https://ross.abutalabs.com/products/openjarvis
Homepage: https://openjarvis.stanford.edu/
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-26T06:11:01+00:00

## Health v2 (maintenance only)
Score: 73/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 73, longevity 14
- inputs: {"age_days": 200, "days_push": 7, "days_rel": 100, "gap_med": 37.0, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 9045, forks 2089 (observed 2026-08-28T04:10:28.071613+00:00)

## What it is
OpenJarvis is a Python framework for building local-first personal AI agents that run on your own hardware, with cloud APIs as an optional fallback. It provides shared primitives for intelligence, agents, tools, memory, and a learning loop, plus evaluations that treat energy, latency, and cost as first-class constraints.

## Use cases
- run a personal AI assistant entirely on my own device
- build local AI agents that only call the cloud when necessary
- serve an OpenAI-compatible API from a local model
- compare local LLMs by energy, latency, and cost on my hardware
- build a coding assistant that reads my repo without sending code to the cloud
- improve a local model from my own interaction traces

## When to choose
- you want privacy-first, on-device AI with zero cloud dependency by default
- you want a unified layer over Ollama, vLLM, SGLang, and llama.cpp
- you care about energy/cost-aware evaluation of local models
- you want MCP tool support and persistent local memory in an agent stack

## When to avoid
- you need maximum-quality frontier model output regardless of cost or privacy
- you want a fully managed hosted service with no local setup
- you need non-Python integration or a polished end-user GUI app

## Facets
- artifact type: framework
- maturity: active
- function: llm-inference, agent-framework, rag, mcp, chatbot, machine-learning, http-server, sdk
- domain: artificial-intelligence, large-language-models, developer-tools, privacy, self-hosted
- platform: windows, python, cli, self-hosted, cross-platform
- tags: local-first, on-device-ai, ollama, vllm, llama-cpp, openai-compatible-api, energy-efficiency, personal-ai, stanford, model-catalog, trace-learning, ai-agents, retrieval-augmented-generation, macos, linux, gpu

## Member repositories
- open-jarvis/OpenJarvis (main) score 73

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:28.071613+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:24:03.874244+00:00, confidence not recorded.
  - readme: https://github.com/open-jarvis/OpenJarvis (fetched 2026-08-28T04:10:28.071613+00:00, sha ea4f492ff2ec)
  - homepage: https://openjarvis.stanford.edu/ (fetched 2026-08-29T08:24:01.640388+00:00, sha fe9d157b528c)
  - registry_pypi: https://pypi.org/pypi/openjarvis/json (fetched 2026-08-29T08:24:01.650944+00:00, sha 5c7b85ab585d)
- Data as of 2026-08-30T08:39:29.467469+00:00.