InferenceIndex

Docs

Install and quickstart

Request access, then swap your base URL and model name once you have access.

A Compiled Agent Model uses the same OpenAI-compatible chat completions request you already send. The client stays the one you use today. Two values change: the base URL and the model name.

Request access

  1. Open https://inferenceindex.com/#waitlist.
  2. Choose Sign up with Google.
  3. Swap your base URL and model name once you have access.

Your Google name, email, and profile photo are saved to the InferenceIndex waitlist for access and product updates.

Once you have access

Swap your base URL and model name once you have access. The rest of the chat completions call stays the same.

  • Python clients set the base URL on base_url.
  • TypeScript clients set the base URL on baseURL.
  • curl posts to the chat completions path on that base URL.

Related

  • What it does and when to use it: https://inferenceindex.com/docs