Leave empty to use the provider's own endpoint. Set it for a local model, for example http://localhost:11434/v1 for Ollama, or to route requests through an AI gateway.

With Cloudflare AI Gateway, use the gateway's provider URL, for example https://gateway.ai.cloudflare.com/v1/ACCOUNT/GATEWAY/anthropic or .../openai. The gateway can then log, rate-limit and scan requests. It stores prompts, which contain Jenkins data, according to your gateway settings.