What is this?
Odysseus is a self-hosted AI workspace that bundles chat, agents, deep research, a document editor, email, notes, tasks, and calendar around local or API models.
Odysseus is a self-hosted AI workspace that bundles chat, agents, deep research, a document editor, email, notes, tasks, and calendar around local or API models.Is this a hosted SaaS or something I run myself?You run it yourself: it's a Dockerized app with its own containers, not a cloud service — you self-host it on your own machine or server.
You run it yourself: it's a Dockerized app with its own containers, not a cloud service — you self-host it on your own machine or server.How do I get it running?Start it with docker compose; once the containers are healthy, open the app in your browser. The first admin password is printed in the docker compose logs for the odysseus service.
Start it with docker compose; once the containers are healthy, open the app in your browser. The first admin password is printed in the docker compose logs for the odysseus service.Do I need Docker, or can I install it natively?Docker is the default path, but native installs, GPU setup, and HTTPS/config details for people who skip containers are covered in the separate setup guide.
Docker is the default path, but native installs, GPU setup, and HTTPS/config details for people who skip containers are covered in the separate setup guide.What's the main thing to watch out for when deploying it?Security is on you: keep AUTHENABLED=true for any network-accessible deployment, keep LOCALHOSTBYPASS=false outside local development, and never expose raw model/service ports publicly.
Security is on you: keep AUTHENABLED=true for any network-accessible deployment, keep LOCALHOSTBYPASS=false outside local development, and never expose raw model/service ports publicly.Does it only run on Linux, or also Windows and macOS?Besides the default Linux/Docker setup, the guide has dedicated Windows and macOS install instructions plus GPU notes.
Besides the default Linux/Docker setup, the guide has dedicated Windows and macOS install instructions plus GPU notes.Since it reads my email and notes, where does that data go?Everything stays on your own self-hosted server; the README explicitly warns to keep private data out of Git and never expose raw model/service ports to the public internet.
Everything stays on your own self-hosted server; the README explicitly warns to keep private data out of Git and never expose raw model/service ports to the public internet.What can the Email feature actually do?It connects to your inbox over IMAP/SMTP and handles triage, tagging, summaries, reminders, and drafting replies.
It connects to your inbox over IMAP/SMTP and handles triage, tagging, summaries, reminders, and drafting replies.What is the Deep Research mode?A multi-step agent that does web research, reads sources, and writes a synthesized report on its own.
A multi-step agent that does web research, reads sources, and writes a synthesized report on its own.What's the Cookbook feature for?It gives hardware-aware local model recommendations along with downloading and serving them directly from the app.
It gives hardware-aware local model recommendations along with downloading and serving them directly from the app.Does it support MCP and external tools?Yes — the chat/agents module can use local or API models together with tools, MCP, file and shell access, skills, and memory.
Yes — the chat/agents module can use local or API models together with tools, MCP, file and shell access, skills, and memory.What are similar projects?mail, stalwart, mailinabox, rowboat, khoj, EnvoyMesh, taskuary, atomic-agent
mail, stalwart, mailinabox, rowboat, khoj, EnvoyMesh, taskuary, atomic-agentWhich topics is it in?imap, smtp, web-search, calendar, 2fa, task-management, note-taking, email, chat, local-llm, agpl, agents
imap, smtp, web-search, calendar, 2fa, task-management, note-taking, email, chat, local-llm, agpl, agents