AI connected to your data, in France
A single engine to query your data and run your models: Spice.ai accelerates SQL, vector search, and LLM inference, installed and maintained by DINAO on our French servers.
What is Spice AI?
Spice.ai OSS is a portable and lightweight runtime (a single binary of about 140 MB) that combines the capabilities of a fast data warehouse and a self-hostable AI inference engine in one deployable package. Written in Rust and released under the Apache 2.0 license, it integrates at the application scale, often deployed as a sidecar.
On the data side, Spice provides a SQL API to query structured and unstructured data across databases, warehouses, and data lakes, and materializes/caches results in DuckDB, SQLite, or PostgreSQL. On the AI side, it offers an OpenAI-compatible API for local or remote inference, vector and hybrid search, text-to-SQL, memory, evaluations, and observability.
It is the ideal tool for building AI applications and agents grounded in your data, with the flexibility of cloud, edge, or on-premise deployment.
Host Spice AI at DINAO
Resource tiers compatible with Spice AI prerequisites (minimum 1 vCPU / 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…
What is Spice.ai used for concretely?
To build AI applications and agents "grounded" in your data: it unifies SQL access to your sources, accelerates queries, and serves your models via an OpenAI-compatible API, all within a single portable runtime.
Do my data go to an AI provider?
In local mode, no: inference and search run on your instance in France. If you connect an external API, your content goes to that provider instead — prefer local mode to remain sovereign.
What data sources can I connect?
Spice exposes a single SQL interface above databases, warehouses, and data lakes, and accelerates results via DuckDB, SQLite, or PostgreSQL, in memory or on disk.
Is a GPU required?
Not necessarily. Small models run on CPU; for heavier inference, a GPU (optional or dedicated) is recommended depending on the target models.
Is Spice open source?
Yes, Spice.ai OSS is under the Apache 2.0 license and deploys as a binary or self-hostable Docker image — this is what DINAO operates for you in France.
