Loved by enterprises
What our customers are saying
Yours for the voyage.
Odysseus is a self-hosted interface for talking to language models — chat, autonomous agents, tools, model serving, email, research, and more. Local-first, privacy-first, and no telemetry. Just you and your models.
(if you want to add an API that's cool too — I'm not here to tell you how to live your life…)
Started as an AI chat. Became a workspace. Each piece runs locally against whatever endpoints you point it at.
Multi-turn chat plus autonomous agents that plan, call tools, and work through tasks.
Built-in tools (bash, files, web, memory) plus any MCP server you connect. Toggle per tool.
Hardware-aware model recommendations and one-click serving across 270+ catalogued models.
AI summaries, style-matched draft replies, auto-tagging and spam triage over IMAP/SMTP.
Multi-step research runs that gather, read, and synthesize sources into a written report.
Send one prompt to several models at once and compare their answers side-by-side.
Persistent memory the assistant builds up and recalls across all your conversations.
The assistant writes, refines, and reuses its own skills — getting more capable over time.
Runs on your machine against your own endpoints. No telemetry, with optional external integrations when you choose them.
Odysseus was created by a carefully crafted one-shot AI prompt:
> idk what to make can you write it for me?
actually make an ai chat, but make it good
and also make it better
Each panel expands and plays its preview when you hover it.
I started working on the Odysseus project because running local AI felt fun — a step into the future. But the options to actually engage with LLMs felt like taking steps back. Where were features like Memory, Deep Research, Agents, and just basic integrations?!
So I started building my own, for fun — and eventually figured it might be fun to share what I built for myself with others. Doesn't work for you? Well… it runs great on my hardware.
It's open source and free. No sales team, no demo request — just clone the repo.