What it is
Odysseus is a self-hosted AI workspace that combines chat and agents, deep research, a document editor, email, notes, tasks, calendar and local model workflows in one app. It supports local and API models, tools, MCP, files, shell, skills and memory. The quick start runs it with Docker Compose, using an official multi-arch image from ghcr.io. It is licensed AGPL-3.0-or-later.
Who it's for
- Users who want a self-hosted AI workspace instead of a hosted service
- People running local or API models who want chat, agents and tools in one interface
- Self-hosters comfortable deploying with Docker Compose or Portainer
- Contributors interested in fresh-install testing, provider setup bugs, mobile/editor polish and docs
Requirements
Requirements
- Docker with Docker Compose (the quick start uses docker compose)
- Git, to clone the repository
- A .env file copied from .env.example
- Port 7011 on localhost for the web UI
Setup
Clone, configure and start
Clone the repo, copy the example env file, and start the containers. The compose files pull the official image and only build locally if the pull fails.
bashgit clone https://github.com/odysseus-dev/odysseus.git cd odysseus cp .env.example .env docker compose up -d --buildOpen the app and get the admin password
Open http://localhost:7011 when the containers are healthy. The first admin password is printed in the output of
docker compose logs odysseus.Pin an immutable image for production
For production, pin the immutable :X.Y.Z-<sha> tag instead of :latest, which moves on every push to main.
bashODYSSEUS_IMAGE=ghcr.io/odysseus-dev/odysseus:1.0.2-7c8070f docker compose up -d
Examples
Quick start with Docker Compose
bashgit clone https://github.com/odysseus-dev/odysseus.git
cd odysseus
cp .env.example .env
docker compose up -d --buildWhat it does: Clones the repository, creates the environment file and starts Odysseus in the background.
Pin a specific build
bashODYSSEUS_IMAGE=ghcr.io/odysseus-dev/odysseus:1.0.2-7c8070f docker compose up -dWhat it does: Overrides the image with an immutable tag so production deployments always run one specific build.
Pros & cons
Pros
- Pro:Broad feature set in one workspace: chat, agents, deep research, documents, email, notes, tasks and calendar
- Pro:Supports both local and API models, plus MCP, tools, skills and memory
- Pro:Simple Docker Compose deployment with an official multi-arch image and immutable tags for pinning
- Pro:Includes security options such as AUTH_ENABLED and 2FA
Cons
- Con:Has powerful local tools (such as shell access), so it needs careful security configuration and must not expose raw model/service ports publicly
- Con:The default branch dev gets changes first, while main is the more curated branch
- Con:AGPL-3.0-or-later license may constrain some uses
Images
