Design your AI agents, hosted in France
A visual programming environment for generative AI powered by Rivet (Ironclad), installed and maintained by DINAO. Build and debug your prompt chains as graphs — your data stays on our French servers, under your control.
What is Rivet?
Rivet, created by Ironclad, is an open source visual programming environment for generative AI. It allows you to design complex agents and prompt chains by assembling nodes in a graph, rather than coding an orchestrator by hand. Flows can split to run in parallel, then converge again.
Rivet offers a collaborative graphical interface to design, debug, and deploy LLM chains. Remote debugging allows you to connect to an endpoint in your application and visualize the chain's execution in real time. Technically, Rivet core is a TypeScript library that executes the graphs: your application can call your Rivet graphs, and Rivet can call your application's code.
Since graphs are simple YAML files, you version them in your repository and review them like code. Managed by DINAO in France and connected to a local model, Rivet provides you with a complete AI engineering workshop while keeping 100% of your data in France.
Host Rivet at DINAO
Resource tiers compatible with Rivet 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…
What is Rivet used for?
Rivet is a visual environment for designing, debugging, and deploying complex AI agents and prompt chains. You assemble nodes in a graph rather than writing an orchestrator by hand.
Do I need to know how to code?
Not to design the flows: the interface is visual, from no-code to low-code. Your developers can go further by integrating Rivet core (TypeScript) into their applications.
Do my prompts go to OpenAI?
Not in local mode. Connected to an open-weight model hosted on your DINAO instance, your prompts and data remain in France. This is the truly sovereign mode (see the sovereignty box).
Can I version my agents?
Yes. Rivet graphs are simple YAML files: you version them in your Git repository and review them like any other code.
Can I integrate Rivet into my application?
Yes. Rivet core is a TypeScript library: your application can execute your graphs, and your graphs can call your application's code, with real-time remote debugging.
