odysseus: Run a Self-Hosted AI Workspace

Overview

Odysseus is a self-hosted workspace that combines AI chat and agents with research and everyday work tools, including documents, email, notes, tasks, and calendar. It addresses the need to switch between separate AI and productivity apps by bringing those workflows together in one deployment.

It is aimed at individuals and teams comfortable operating their own software and configuring model providers or local models. The repository's dev branch receives changes first; main is described as the more curated branch.

Key Features

• Chat and agents that can use local or API models, tools, MCP, files, shell, skills, and memory.
• Deep Research for multi-step web research, source reading, and report generation.
• Model recommendations, downloads, and serving through the Cookbook.
• Side-by-side model comparisons with blind testing and synthesis.
• A writing-focused document editor with AI edits and support for Markdown, HTML, and CSV.
• Email workflows with IMAP/SMTP, triage, summaries, reminders, and reply drafts.
• Notes, tasks, reminders, scheduled agent tasks, and CalDAV calendar sync.

Use Cases

• A self-hoster who wants AI chat, agents, and productivity tools in one workspace they operate themselves.
• A local-model user who wants to manage model recommendations, downloads, and serving alongside chat and agent workflows.
• A researcher who needs web research, source review, and generated reports in the same environment as their notes and documents.
• A small team or individual handling email and planning who wants AI-assisted triage, drafts, reminders, and calendar workflows.

Getting Started

git clone https://github.com/odysseus-dev/odysseus.git
cd odysseus
cp .env.example .env
docker compose up -d --build

Then open http://localhost:7000 after the containers are healthy. See the README and setup guide for configuration and deployment details.

Alternatives

• QwenPaw: QwenPaw is a personal assistant deployed through chat applications, while Odysseus is a self-hosted workspace with built-in research, documents, email, notes, and calendar tools.

Considerations

• This is a self-hosted application, so deployment, updates, and access controls are the operator's responsibility.
• The project recommends keeping authentication enabled for network-accessible deployments and not exposing raw model or service ports publicly.
• The dev branch gets changes first; use main for the more curated branch, according to the README.
• AGPL-3.0 licensing is an important consideration for users modifying or distributing the software.
• Local model workflows and GPU setup are covered in the setup guide; the README does not say that a GPU is required.