
Replace the OpenAI API with a 100% French instance
Text, images, audio, embeddings : LocalAI is the drop-in replacement for the OpenAI API, installed on French GPUs and maintained by DINAO. A single private API for all generative AI — without sending any data overseas.
What is LocalAI?
LocalAI is a free and local alternative to OpenAI and Anthropic: a "small composable AI stack" that runs language models, agents, and document intelligence on your own hardware, without cloud dependency. It runs on consumer-grade hardware and does not require a GPU.
Its main advantage: it is a direct (drop-in) replacement for the OpenAI API. Applications and libraries already written for OpenAI work simply by pointing to LocalAI, while remaining 100% local. It supports LLM inference, image and audio generation, multiple model families, and multiple backends.
LocalAI integrates with LocalAGI (autonomous no-code agents) and LocalRecall (memory and semantic search). Installation is available via Docker, Kubernetes, or binaries. The project is open source under the MIT license.
Host LocalAI at DINAO
Resource tiers compatible with LocalAI prerequisites (minimum 4 cœurs / 8 Go / 20 Go). Hosted in France, fully managed.
- 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
2 hidden tier(s) (insufficient resources for this app) : Découverte, Standard
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 can LocalAI do?
LocalAI covers all generative AI : text generation, images (Stable Diffusion), audio transcription and synthesis (Whisper, TTS), and embeddings. All via a single OpenAI-compatible API.
Will my OpenAI applications work?
Yes. LocalAI is a drop-in replacement: you change the base URL and key, and your /chat, /images, /audio, and /embeddings calls work without modifying your code.
Where is the data hosted?
On the DINAO GPU infrastructure in France, in one of the available datacenters. Your queries and generated content do not leave the territory.
Do I need technical skills?
To integrate the API, some development knowledge is useful. DINAO manages installation, GPU, models, and updates. You can also connect a graphical interface on top.
Can I change plans or models?
Yes. You can upgrade at any time and install models from the gallery according to your needs — no proprietary lock-in.