KEDA / GUIDES
Connect to the Keda OpenAI-compatible API
Create an API key, configure your base URL, select a model ID and verify your first request and streamed response.
1. Create a key and check access
Sign in to the developer console and create a key under APIs & Routing. Check the workspace, available credit and model permissions. Store the key securely. Models & Pricing shows the catalog; actual access depends on your key configuration.
2. Configure the endpoint and model
Set your OpenAI-compatible client base_url to https://keda8.ai/v1 and use the key created in the Keda console. Set model to an exact catalog ID. Avoid duplicating /v1. Verify support for Chat Completions, Responses or other endpoints individually.
3. Verify the response and usage
Start with a short text request and inspect the HTTP status, response body and usage records. For 401 check the key; for 403 check permissions; for 429 check quota and rate limits. Streaming clients must consume SSE events correctly. See the API docs for examples and error handling.