Your AI second brain, hosted in France
A personal and self-hosted AI assistant powered by Khoj, installed and maintained by DINAO. Query your documents and the web, create agents — your notes and research stay on our French servers.
What is Khoj?
Khoj is a personal open-source and self-hostable AI assistant, often described as a "second brain." It combines semantic search on your private documents with conversational AI to answer your questions based on your own knowledge as well as the web.
Khoj indexes a wide variety of sources — PDF, Markdown, Word, Notion, org-mode, images — and retrieves relevant information by meaning rather than just keywords. Beyond search, it allows you to create agents with personas, knowledge, and dedicated tools, schedule automations, and conduct in-depth web research (deep research) thanks to the integration of SearxNG and a code execution sandbox.
Compatible with any LLM — local open-weight models via Ollama or cloud APIs (OpenAI, Anthropic, Gemini, DeepSeek) — Khoj can be used from the browser, Obsidian, Emacs, the desktop app, or mobile. Backed by Y Combinator and deployable in containers, it offers sovereign document AI when connected to local models.
Host Khoj at DINAO
Resource tiers compatible with Khoj prerequisites (minimum 2 cœurs / 4 Go / 10 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 is an "AI second brain"?
Khoj indexes your notes and documents, then answers your questions based on their content, like an assistant that has read everything. It combines semantic search and conversational AI to retrieve and synthesize your knowledge.
Are my documents used to train models?
No. In local mode, your notes and conversations remain private on your instance and are never shared or used for training, unlike public services.
Can I use my own models?
Yes. Khoj works with locally hosted open-weight models (via Ollama) or with an external API of your choice. For total sovereignty, prefer local models.
Where can I access it from?
From your browser, but also via Obsidian, Emacs integrations, the desktop app, and mobile, all connected to your instance hosted in France.
Where is the data hosted?
On DINAO infrastructure in France, in one of the available datacenters. Your documents and searches do not leave the territory in local mode.
