Build high-quality AI datasets in France
An annotation and curation platform powered by Argilla, installed and maintained by DINAO. To ensure the reliability of your AI datasets — your data stays on our French servers.
What is Argilla?
Argilla is an open source data curation tool that allows you to annotate and review datasets, then publish curated datasets (for example, to the Hugging Face Hub). It is a collaboration platform designed to bring together AI engineers and domain experts to produce high-quality data.
Argilla is used to collect human feedback for a wide variety of projects: traditional NLP (text classification, NER…), LLMs (RAG, preference alignment…), or multimodal models (text-to-image…). Its programmatic approach enables the creation of continuous evaluation and model improvement workflows.
The goal: ensure that data work yields results by iterating quickly on the right data and models. Argilla also integrates with LLMs for pre-annotation and synthetic data generation.
Host Argilla at DINAO
Resource tiers compatible with Argilla prerequisites (minimum 2 vCPU / 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 Argilla used for?
To build high-quality datasets for AI: annotate and review data, collect human feedback for NLP, LLMs (RAG, preference alignment) or multimodal models.
Can I use an LLM for pre-annotation?
Yes. Argilla integrates with LLMs to pre-annotate or generate synthetic data. By hosting a local model on your DINAO instance, you keep these processes in France.
Where is the data hosted?
On DINAO infrastructure in France, in one of the available data centers. Your data and annotations never leave the country.
Are technical skills required?
Annotation via the interface is accessible to domain experts. Setting up workflows and the Python SDK is handled by a data team. DINAO manages installation, security, and updates.
Can I export my datasets?
Yes. Your datasets can be exported to your pipelines or the Hugging Face Hub — no proprietary lock-in.
