Skip to main content
Applies to:


Summary Goal: Override organization-level custom AI providers at the project level to use different API keys or endpoints per project. Features: Project-level AI providers, custom provider configuration, provider name matching for override behavior.

​
Configuration Steps

​
Step 1: Create organization-level custom provider

Set up your default custom provider at the organization level.
  1. Go to Settings → Organization → AI providers
  2. Under Custom providers , click Create
  3. Configure your provider (e.g., name: openai-custom, endpoint, API key)
  4. Save the provider

​
Step 2: Create project-level override

Create a project-level custom provider with the exact same name to override the organization-level provider.
  1. Go to Settings → Project → Project-level AI providers
  2. Under Custom providers , click Create
  3. Use the exact same provider name as the org-level provider you want to override
  4. Configure with your project-specific settings (e.g., different API key, regional endpoint)
  5. Save the provider

​
Step 3: Verify override behavior

The project will now use the project-level provider instead of the organization-level provider when names match.
  • Project-level providers with matching names completely override org-level providers
  • Other org-level providers remain accessible to the project
  • Use this for regional API keys (US vs non-US OpenAI endpoints) or project-specific configurations

​
Common Use Cases

  • Regional endpoints: Override org-level non-US OpenAI keys with US-specific keys at project level
  • Project isolation: Provide project-specific API keys while maintaining org-level defaults
  • Custom models: Override default model endpoints for specific projects

​
Key Requirements

  • Provider names must match exactly (case-sensitive) for override to work
  • Project-level providers take precedence when names match
  • Non-matching org-level providers remain accessible to projects