Jev Lint — npmjs.com

Nothing needs installing to try it: set TYPESAFE_API_KEY and run npx -y jev-lint check examples with Node 24+ and npx. For a project, npm install --save-dev jev-lint, then npx jev-lint init writes a .jev-lint.yaml and npx jev-lint check src --dry-run shows what it would ask and the price.

A rule is one sentence, such as the comment above this code claims something that is not true of the code. An ast-grep matcher selects which code the model is asked about, and the answer is a probability; an answer at or above the rule's threshold becomes a finding.

Yes. The code being judged is sent to an API, which is what the key pays for; nothing runs locally except the ast-grep matcher. The key is read from the environment only and never from the config file, which belongs in version control.

jev-lint reads the key from the environment only, as TYPESAFE_API_KEY or TYPESAFEAI_API_KEY. The config file holds rules, thresholds and file selections, not a key, so it can be committed.

Everything that does not ask the model works without a key: --dry-run prices a run, jev-lint rules lists every loaded rule with its question, cutoff and file, jev-lint replay re-scores a recorded run, and CI can lint from a committed cache with no key at all.

It is not for anything a compiler, type checker or ESLint already decides. It is good at code that contradicts itself and poor at defects needing knowledge of a specific API, such as .sort() defaulting to lexicographic order; keep existing tools for those.

The documented run reports six findings over two files for a tenth of a cent, with the cost of a bigger run measured rather than extrapolated in docs/cost.md. --dry-run reports the price of a run without making a request.