Install

jetlint is in active development. From v0.1.0 onward, prebuilt binaries ship via npm; until then, install from source with Go.

Install via npm (v0.1.0 and later)

The wrapper package is @jetlint/cli. It pulls in the right prebuilt binary for your platform automatically via optionalDependencies.

# npm
npm install --save-dev @jetlint/cli

# pnpm
pnpm add -D @jetlint/cli

# yarn
yarn add --dev @jetlint/cli

# bun
bun add -d @jetlint/cli

The installed CLI is jetlint, regardless of package manager:

npx jetlint --project ./tsconfig.json

Supported platforms: linux x64, linux arm64, macOS x64, macOS arm64, Windows x64. On anything else, fall back to installing from source.

Install from source

This is the current path while v0.1.0 is in progress, and the fallback for platforms without a prebuilt binary.

Requirements

From source

go install github.com/jetlint/jetlint/cmd/jetlint@latest

This drops a jetlint binary into $(go env GOPATH)/bin. Make sure that directory is on your PATH.

Run on a project

jetlint --project ./tsconfig.json

By default, jetlint applies the 5 recommended rules at error severity. The other 177 rules ship off; see Config to opt in.

Run as a daemon

jetlint runs as a long-lived sidecar to keep the TypeScript checker warm. Subsequent lints reuse the program and complete in milliseconds.

# start the daemon (logs to stderr; reads JSON-RPC on stdin)
jetlint daemon --project ./tsconfig.json

The daemon speaks JSON-RPC over stdio; editor extensions and CI runners can attach without paying the program-load cost on every request.

Editor integration

Editor packages are not yet published. The daemon’s JSON-RPC schema is stable enough to script against for early integration work; see the issue tracker for design notes and the in-progress LSP bridge.

CI

Run jetlint as a single command in your pipeline. Exit codes:

CodeMeaning
0No diagnostics
1Lint diagnostics reported
2Configuration or invocation error
jetlint --project ./tsconfig.json --format=github

--format=github emits GitHub Actions workflow commands so diagnostics surface inline on pull requests.