
Your personal AI assistant, hosted in France
Open Agent brings together LLMs, your knowledge base, and autonomous agent loops in a self-hostable platform. Installed and maintained by DINAO, your data remains yours, on our French servers.
What is Open Agent?
Open Agent is a new-generation open-source, self-hostable personal AI assistant. It brings together powerful language models, your own knowledge base, and autonomous agent loops within a single platform. You connect your chosen model provider, build a RAG from your documents, and let the agent act on your behalf.
Beyond dialogue, Open Agent supports computer-use, browser-use, and code agent capabilities: it can browse the web, execute code, and call any MCP (Model Context Protocol) compatible tool. You can connect any major LLM provider and switch from one to another conversation by conversation, without modifying the code.
Fully open source and designed for self-hosting, Open Agent creates no proprietary lock-in: your data remains yours, always. Available as pre-compiled binaries for Linux, macOS, and Windows, it integrates naturally into infrastructure managed by DINAO.
Host Open Agent at DINAO
Resource tiers compatible with Open Agent prerequisites (minimum 2 vCPU / 2 Go / 5 Go). Hosted in France, fully managed.
- 2 dedicated vCPU
- 4 Go RAM
- 40 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 4 dedicated vCPU
- 8 Go RAM
- 80 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 8 dedicated vCPU
- 16 Go RAM
- 160 GB NVMe
- Daily backups
- Managed & monitored by DINAO
1 hidden tier(s) (insufficient resources for this app) : Découverte
This 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…
What distinguishes Open Agent from a chatbot?
Open Agent does more than just answer: thanks to agent loops, it can browse the web, execute code, and call your tools via MCP to accomplish concrete tasks, relying on your RAG knowledge base.
Is my data used to train models?
No. In local mode, your documents and conversations remain on your instance in France. Your data remains yours and is not shared.
Which models can I connect?
Any major LLM provider, locally (open-weight model served on your instance) or via external API with your key. You can switch models per conversation.
What does MCP bring?
The Model Context Protocol allows the agent to call standardized tools (files, APIs, browser, scripts). You thus extend the agent's capabilities to your own internal systems.
Can I change plans or export my data?
Yes. Open Agent is lock-in free: you can upgrade or downgrade at any time, and your data remains exportable.