Odysseus

Odysseus brings AI chat, autonomous agents and personal work into one workspace on your own hardware. It's for people who want models to work with their documents and ongoing tasks in a self-hosted app.

• Chat and agents use local models or external APIs, with persistent conversation memory, reusable agent-written skills, shell and file tools, and optional MCP connections you can control per tool.
• Local model management recommends models based on your hardware and handles downloads and serving across machines; comparison tools show responses to the same prompt side by side or hide model identities for blind testing.
• Web research searches and reads sources across multiple steps, then writes a cited report.
• Documents and images support AI edits and writing suggestions in a document editor, plus image generation, editing, background removal and inpainting in a gallery.
• Email and planning summarize, tag and triage IMAP/SMTP mail and draft replies in your style, alongside notes, reminders, to-dos, scheduled agent tasks and CalDAV calendar sync.

Odysseus is free and open source under AGPL-3.0 and runs in Docker. The workspace runs locally against the model endpoints you choose and collects no telemetry. External model APIs and integrations send the relevant data to those services, and web research requires access to online sources.