Odysseus is a self-hosted AI workspace that bundles chat, agents, research, documents, email, notes, calendar, and local model workflows into one application. Its GitHub repository describes it as “a self-hosted AI workspace for chat, agents, research, documents, email, notes, calendar, and local model workflows.”
The pitch is straightforward: instead of spreading AI chat, deep research, document editing, and personal productivity across several cloud services, Odysseus runs all of it on hardware you control. The project is written mostly in Python with a JavaScript frontend, and it is released under the AGPL-3.0-or-later license.
Why this project matters
Most AI workspaces are cloud products. That is convenient, but it means your conversations, documents, email, and notes live on someone else’s infrastructure, and your model choices are limited to what the vendor offers.
Odysseus takes the opposite approach. It is designed to run on your own server or workstation, work with local models as well as API models, and keep the surrounding productivity data — mail, tasks, calendar entries, documents — in the same self-hosted stack. For homelab users and privacy-conscious teams, that combination is the main draw.
What is included
The repository lists a broad feature set that goes well beyond a chat UI:
- Chat and agents with support for local and API models, tools, MCP integration, file handling, shell access, skills, and memory.
- Cookbook, which recommends models based on your hardware and handles downloading and serving them.
- Deep research for multi-step web research with source reading and report generation.
- Comparison tools for blind testing models and synthesizing their answers.
- A document editor focused on writing, with AI edits and suggestions plus Markdown, HTML, CSV, and code highlighting.
- Email over IMAP/SMTP with inbox triage, tagging, summaries, and draft generation.
- Productivity features such as notes, reminders, tasks, scheduled agent runs, and CalDAV synchronization.
- Extras including an image gallery and editor, theming, file uploads, web search, and two-factor authentication.
That scope makes Odysseus closer to a personal AI operating environment than a single-purpose tool. Whether you want all of it is a fair question, but the parts are integrated by design rather than stitched together from separate services.
Getting started with Docker
The documented quick start uses Docker Compose:
git clone https://github.com/pewdiepie-archdaemon/odysseus.git
cd odysseus
cp .env.example .env
docker compose up -d --build
Once the containers are healthy, the application is available at http://localhost:7000, and the initial admin password is printed in the compose logs.
The project also ships GPU-oriented compose variants for NVIDIA and AMD hardware, which matters if you plan to serve local models rather than relying on API providers. The setup guide in the repository covers native installations, GPU configuration, Windows and macOS directions, HTTPS setup, and further configuration details.
Practical considerations
A few points are worth knowing before you deploy it:
- Hardware expectations depend on your model plans. Running Odysseus purely against API models is a very different footprint from serving large local models. The built-in hardware-aware model recommendations are there to help with exactly that decision.
- Branch strategy. The
devbranch receives updates first, whilemainis the more stable option. For a production-ish home server,mainis the sensible default. - License. AGPL-3.0-or-later is a strong copyleft license. That is irrelevant for personal self-hosting, but organizations embedding it in a service should review the obligations.
- Email integration means credentials. Connecting IMAP/SMTP accounts to any self-hosted tool puts real credentials in your stack, so the usual care around
.envhandling, HTTPS, and access control applies. Two-factor authentication is available in the app itself.
Contributing
The project is actively developed and welcomes contributions, with a stated focus on testing, bug fixes, documentation, and focused refactors. The repository includes setup guides for Linux, macOS, and Windows, a security policy, and a roadmap.
Verdict
Odysseus is an ambitious take on the self-hosted AI workspace: one deployment that covers chat, agents, research, documents, email, and personal organization, with first-class support for local models. If you already run a homelab and want your AI workflows under your own roof, it is worth a look — start with the Docker quick start on the main branch and grow into the GPU and email features as needed.
