open-jarvis/OpenJarvis

Personal AI, On Personal Devices

What it solves

OpenJarvis addresses the dependency of personal AI agents on cloud APIs, which creates privacy and reliability concerns. It provides a software stack that allows users to run powerful personal AI agents locally on their own devices, reducing reliance on external servers and optimizing for on-device constraints like energy, latency, and cost.

How it works

OpenJarvis is a framework for local-first AI built on three pillars: shared primitives for building on-device agents, a learning loop that uses local trace data to improve models, and an evaluation system that treats hardware constraints (FLOPs, energy, and dollar cost) as primary metrics alongside accuracy. It integrates with tools like Ollama for model serving and supports a wide variety of "skills" (tools) that agents can discover and invoke via the agentskills.io standard.

Who it’s for

It is designed for developers and researchers who want to build and deploy autonomous personal AI agents that run locally by default, as well as users who want a private, on-device AI assistant for tasks like research, coding, and daily scheduling.

Highlights

  • Local-First Execution: Designed to run on personal devices, calling the cloud only when necessary.
  • Diverse Agent Modes: Includes built-in agents for scheduled tasks (e.g., morning briefings), continuous monitoring, and on-demand reasoning.
  • Extensible Skill System: Supports importing thousands of community skills from sources like Hermes Agent and OpenClaw.
  • Hardware-Aware Eval: Evaluates performance based on energy efficiency and latency in addition to accuracy.
  • Cross-Platform Support: Provides easy installation for macOS, Linux, Windows, and desktop GUI installers.

Related

  • Dispatch
    OpenJarvis 1.0 Release: Local-First Personal AI via Ollama

    OpenJarvis 1.0 is an open-source framework for building local-first personal AI agents that integrates with Ollama to run models on personal hardware.

  • Project
    isair/jarvis

    A private, local-first voice assistant that runs on your own hardware, allowing for natural, voice-driven conversations with personal memory and local AI.

  • Project
    elizaOS/eliza

    An open-source TypeScript framework and product stack for building autonomous AI agents that can run locally or in the cloud across web, mobile, and OS distributions.

  • Project
    Rose22/openlumara

    A modular, local-first AI agent framework designed for token efficiency and personal life management, supporting multiple interfaces and sandboxed shell execution.

  • Project
    sandbaseai/sandbase-harness

    SandBase Harness is a self‑hosted runtime that adds persistent sessions, sandboxed tool execution, credential vaults, audit trails and a web console to AI agents. It implements a Claude‑Managed‑Agents‑compatible `/v1` API, stores everything locally in SQLite, supports Docker/K8s sandboxes, works with many model providers, and ships a TypeScript SDK for easy client integration.