Your private AI interface, hosted in France
A ChatGPT-style UI powered by NextChat, installed and maintained by DINAO. Connected to your own models, your conversations stay on our French servers — never with a third-party provider.
What is NextChat?
NextChat (formerly ChatGPT-Next-Web) is a cross-platform chat interface for large language models. Unlike an AI engine, NextChat does not "contain" any models: it is a lightweight presentation layer that connects to an LLM provider via its API. This gives you a "ChatGPT-like" experience (web, installable PWA, native Windows/macOS/Linux apps) while keeping control of the API key and hosting.
For conversations, NextChat handles full Markdown rendering with LaTeX formulas, Mermaid diagrams, and code syntax highlighting, streaming responses, as well as automatic history compression for long conversations. The masks system (prompt templates) allows you to create, share, and debug pre-configured assistants, and plugin support extends its capabilities.
NextChat is compatible with a wide range of providers: OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Azure OpenAI, as well as several Chinese models. A key point for a sovereign host: NextChat is explicitly compatible with self-hosted LLMs and connects to local runtimes (LocalAI, Ollama, vLLM) exposing open-weight models via an OpenAI-compatible API.
Technically, NextChat is written in TypeScript / Next.js, with desktop apps built on Tauri. It requires no database: history is stored in the browser. Its footprint is minimal, making it an ideal UI to connect to a model hosted on DINAO infrastructure.
Host NextChat at DINAO
Resource tiers compatible with NextChat prerequisites (minimum 1 cœur / 512 Mo / 2 Go). Hosted in France, fully managed.
- 1 dedicated vCPU
- 2 Go RAM
- 20 GB NVMe
- Daily backups
- Managed & monitored by DINAO
- 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
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…
Is NextChat an AI model?
No. NextChat is an interface (a client) that connects to a language model via API; it performs no inference itself. It is the model connected behind it that produces the responses.
Do my conversations go to a third party?
History is stored locally in your browser. Only requests sent to the model transit to the configured backend: in local mode (sovereign), they never leave the DINAO infrastructure in France.
Can I use a French/local model without OpenAI?
Yes. NextChat connects to any self-hosted model exposing an OpenAI-compatible API (via LocalAI, Ollama, vLLM…), with open-weight models like Llama, Mistral, or Qwen hosted on your DINAO instance.
Which providers are supported?
OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Azure OpenAI, as well as several Chinese models — and above all any local model exposing an OpenAI-compatible API.
Can I change my plan or model?
Yes. You can upgrade or downgrade at any time and change models on demand. Your data remains exportable — no proprietary lock-in.
