Free tool

LLM savings audit

Drop in a usage export from OpenRouter, OpenAI or any other platform. See what the same tokens cost on LLM Gateway, model by model.

The file is read in your browser and never uploaded.

How the audit works

  1. Your CSV is read row by row in the browser and summed per model.
  2. Model names are matched to the catalogue, ignoring vendor prefixes and dated snapshot suffixes.
  3. Each model is priced at the cheapest provider serving it today, with cached input billed at the cached rate.
  4. The total, including the platform fee, is set against what your file says you paid.

FAQ

Savings audit, answered

Can't find an answer? Contact us

Any CSV with one column for the model and columns for input and output tokens. Activity and usage exports from most LLM platforms work as they are. If a column is not recognized, pick it by hand after loading the file.