We believe AI should be open, private, and accessible to everyone. Jan is building the infrastructure for a future where you own your AI — running locally on your hardware or self-hosted on your servers.

Our mission is to empower individuals, teams, and organizations to harness the power of AI without sacrificing privacy or control.

Open-source ChatGPT alternative that runs 100% offline

Turn your computer into an AI workstation. Jan is a desktop application that lets you run large language models locally with complete privacy.

• Run Local Models — Download and run Llama, Gemma, Mistral, Qwen, and thousands more from HuggingFace
• Connect Cloud Providers — Integrate with OpenAI, Anthropic Claude, Google Gemini, and others
• OpenAI-Compatible API — Local server at localhost:1337 for seamless integration with your tools
• MCP Support — Model Context Protocol for agentic AI capabilities
• Cross-Platform — Available on Windows, macOS, and Linux

Self-hostable, enterprise-grade LLM API platform

Deploy your own AI infrastructure with a single docker compose up. Jan Server is a microservices platform designed for teams and organizations.

• OpenAI-Compatible API — Drop-in replacement for OpenAI endpoints
• Multi-Step Tool Orchestration — Response API for complex agentic workflows
• MCP Tools Integration — Web search, code execution, and custom tools
• Enterprise Auth — OAuth/OIDC via Keycloak with guest access support
• Full Observability — OpenTelemetry, Prometheus, Jaeger, and Grafana built-in
• Media Management — S3 integration for files and attachments

# Get started in minutes
git clone https://github.com/janhq/server
cd server && make quickstart

We're building in the open and welcome contributors of all skill levels.

• 📖 Contributing Guidelines — How to contribute
• 🐛 Report Issues — Found a bug? Let us know
• 💡 Feature Requests — Share your ideas
• 🔒 Security Policy — Report vulnerabilities responsibly

Website • Documentation • Discord • Twitter

_{Built with ❤️ by the Jan team and our amazing community}