Developers have released Odysseus 1.0, a self-hosted AI workspace offering privacy-first alternatives to ChatGPT and Claude. The open-source platform provides chat with local models, AI agents with tool integration, document editing with AI assistance, email triage, and a mobile-responsive interface. Users control their data with local-first architecture.

Installation is straightforward via Docker or manual setup on Linux, macOS, and Windows. The Cookbook feature automatically scans hardware and recommends optimal models, while the Deep Research tool conducts multi-step analysis. The platform supports multiple LLM backends including vLLM, llama.cpp, and OpenAI APIs.

Security-focused design emphasizes proper deployment practices, including HTTPS configuration for network access. The project maintains strict privilege controls, with admin-only access to sensitive functions like shell commands and API tokens. This approach addresses growing concerns about data privacy in AI applications.