Flux RouterDocumentation

Routing and models

How flux-auto picks a model, the three lanes, the receipt headers on every response, and where to find the live model catalogue.

This section explains what happens between your request and the answer: which model Flux picks, which lane it bills, and how you can see both on every response.

If you only want to make your first call, start with Get your key. If you are choosing between letting Flux pick and pinning a model, read Let Flux pick, or pin a model.

In this section

  1. How routing works: what flux-auto does, how it picks among vetted models, and how you stay in control.
  2. Lanes: the Fast, Standard and Reasoning lanes and the aliases that choose them.
  3. Receipts: the headers that show which model answered and what the request cost.
  4. Models: the live catalogue and the ids you can send.

For money, see How billing works and Pricing.