OpenJarvis: Personal AI, on Personal Devices

12 hours ago

OpenJarvis is an open-source framework for personal AI agents that runs entirely on-device.
It provides shared primitives for building on-device agents, efficiency-aware evaluations, and a learning loop that improves models using local trace data.
The framework aims to make personal AI agents run locally by default, calling the cloud only when necessary.
OpenJarvis addresses three missing pieces in today's local AI systems: shared abstractions, efficiency-aware evaluations, and closed-loop optimization.
The framework is structured around five composable primitives: Intelligence, Engine, Agents, Tools & Memory, and Learning.
OpenJarvis treats efficiency as a first-class metric, tracking energy, dollar cost, FLOPs, latency, and related system metrics alongside accuracy.
The framework supports interaction via CLI, browser dashboard, desktop app, and 26+ messaging channels.
OpenJarvis can be used for personal AI tasks, traditional LM workloads, and agentic and long-horizon tasks.
The project is open-source under Apache 2.0 and encourages involvement from researchers, developers, and users.

Hasty Briefsbeta