odysseus

odysseus is a self-hosted AI workspace published by the GitHub organization odysseus-dev. The README describes it as "A self-hosted AI workspace for chat, agents, research, documents, email, notes, calendar, and local model workflows." Once installed, you open local port 7011 in a browser and find chat, agents, documents, email, notes, and calendar in one interface. Everything runs on your own machine, and you can pick a local model or connect your own API account. It also supports MCP, scheduled tasks, and reading uploaded files. The official docs cover installation on Windows, macOS, and GPU setups.

Who it's for:People comfortable with Docker who want their AI assistant, email, calendar, and documents on their own server. Readers with no server experience will find setup hard going.

Before you start: Install with Docker Compose or natively; the first admin password appears in the startup logs. Instances have admin login and two-factor authentication, and the README says to keep authentication on and never expose model ports directly to the internet. odysseus can send and receive email over IMAP and SMTP, read and write CalDAV calendars, process files you upload, and run scheduled tasks. If you choose a cloud model, the relevant content goes to the model service you configure. The README only mentions drafting email replies, and the check did not cover every connector, so AI Inspiration World does not recommend letting it manage an important work inbox directly. The project lists no fixed price; model costs depend on the service you choose. The license is AGPL-3.0-or-later, so running it as a network service comes with source-code obligations.

Based on the project's public documentation. Not tested by AI Inspiration World.