Cost

LLM costs can add up quickly.

Two-pass linting helps to significantly reduce costs by using a cheaper, weaker model for 95% of the work, but if you’re running the linter on very large codebases, LLM costs can still add up quickly.

Every time you run gptlint, it will log the total cost of all LLM calls for that run (if you’re using a supported provider).

Note that this variable cost goes away when using a local LLM, where you’re paying directly for GPU compute instead of paying per token. For most projects, the cost of running gptlint will be orders of magnitude cheaper than relying on your senior engineers to track and fix technical debt.