Your AI workspace, hosted in France
Odysseus brings together chat, autonomous agents, deep research, document editing, and persistent memory in a single self-hosted AI application. Installed and maintained by DINAO, your data stays on our French servers.
What is Odysseus?
Odysseus is a self-hosted AI workspace, designed as the private equivalent of the experience offered by ChatGPT or Claude. Local-first and privacy-oriented, it runs on your own hardware with your own data. It brings together in a single web application chat, autonomous agents, deep research, document editing, notes, and persistent memory.
You can chat with any local or remote model (vLLM, llama.cpp, Ollama, OpenRouter, OpenAI) and assign tools to agents to execute end-to-end tasks. Its deep research feature chains multiple web search steps, reads sources, and produces a documented report.
Regarding tooling, Odysseus offers controlled shell access, file upload, model download, email and calendar integrations, as well as API tokens. It runs on Linux, macOS, and Windows via Docker or native Python, and installs as a PWA on mobile.
Host Odysseus at DINAO
Resource tiers compatible with Odysseus prerequisites (minimum / / ). Hosted in France, fully managed.
This application uses an unofficial/unverified image or requires custom resources. We review and deploy it upon request.
Request a quoteThis application uses AI
The container hosts the application, not the AI engine (which requires dedicated GPUs) : inference is handled externally, with your own provider key. Prioritize a sovereign engine — Mistral AI, NumSpot, Scaleway, or OVHcloud AI Endpoints (France, GDPR) ; an international provider (OpenAI, Anthropic…) only if a specific capability requires it. Inference subscriptions are not included in hosting.
Technical details
You might be wondering…
How does Odysseus differ from a simple AI chat?
Odysseus goes beyond chat: it combines autonomous agents, deep research, document editing, notes, and persistent memory in a single workspace. Agents can chain multiple steps to complete a full task.
Does my data leave the server?
In local mode (model served on your instance), no: everything stays in France on your server. If you connect an external API, the transmitted content goes to that provider.
Which models can I use?
Any local model via vLLM, llama.cpp, or Ollama, or a compatible API (OpenRouter, OpenAI). You choose based on your sovereignty and performance needs.
Do I need technical skills?
No for usage: the interface is ready to use and DINAO handles installation, security, and updates. Technical profiles can additionally leverage advanced tools and agents.
Can I change plans or export my data?
Yes. You can upgrade or downgrade at any time, and your data remains exportable — no proprietary lock-in.
