Teams and enterprise
Your own AI provider
Use your organization's own AI provider and key for Brunel's reasoning, OpenAI, Claude, Mistral or any OpenAI-compatible endpoint, with the key tested, sealed and never shown again.
On Enterprise, the team's owner or admins can set the AI provider Brunel uses for the whole team, in Platform → AI provider.
Supported providers
OpenAI, Claude, Mistral, or any endpoint compatible with the OpenAI chat completions API (for example a model hosted in your own cloud account). Enter the base URL when it is not a standard one, the model name and the key.
What happens to the key
- Brunel calls the provider once to test the key and the model before saving.
- The key is sealed with AES-256-GCM, bound to your team, and never shown again, to anyone.
- The base URL must be a public HTTPS address; calls go through Brunel's outgoing-address checks.
- You can replace or remove it at any time; removing it returns the team to Brunel's default reasoning.
What does not change
Your model gets the same inputs as Brunel's default: facts and short secret-free excerpts, wrapped as untrusted data, with no tools. It still never writes prices, service names, SKUs or Terraform: its output is a schema-validated blueprint that goes through Brunel's mapper, pricer and validator. Monthly token usage is shown in the console.