Run SKILL.md agents, build an enterprise knowledge base with RAG, and deploy a private AI chat box — all on your own infrastructure. No call-home. No data leaving your network.
Drop a SKILL.md into any workspace. OE wires it to your connectors and executes it via the built-in runtime. The same skill runs unchanged on Claude Code, Cursor, Windsurf, or the CLI.
Ingest documents (24+ formats), chunk and embed them into your vector database of choice, and retrieve with cited answers. All on your infrastructure — no data leaving your network.
A fully private AI assistant backed by your own LLM, your own knowledge base, and your own connectors. Embed it as a website widget or use it internally across workspaces.
Docker (recommended)
Open http://localhost:3000 and log in with your super admin credentials. Configure your LLM provider, embedding model, and vector database from the admin panel — no .env changes needed.
Or run from source (Node.js 18+, Yarn 1.x)
Open http://localhost:<FRONTEND_PORT> — the port is set by FRONTEND_PORT in server/.env.
All configurable from the admin panel — no .env changes, no code.
No usage limits. No telemetry. No call-home. Use, modify, and deploy for any purpose.
Self-hosted. Apache-2.0. No call-home. Own your data completely.