Use LLMs to enforce best practices across your codebase.

GPTLint

A fundamentally new approach to code quality. Use LLMs to enforce higher-level best practices across your codebase in a way that takes traditional static analysis tools like eslint to the next level.

Features

• ✅️ enforce higher-level best practices that are impossible with ast-based approaches
• ✅️ simple markdown format for rules (example, spec)
• ✅️ easy to disable or customize rules
• ✅️ add custom, project-specific rules
• ✅️ same cli and config format as eslint
• ✅️ supports gptlint.config.js and inline overrides /* gptlint-disable */
• ✅️ content-based caching
• ✅️ outputs LLM stats per run (cost, tokens, etc)
• ✅️ built-in rules are extensively tested w/ evals
• ✅️ supports all major LLM providers and local models
• ✅️ augments eslint instead of trying to replace it (we love eslint!)
• ✅️ includes guidelines for creating your own rules
• ❌ MVP rules are JS/TS only for now
• ❌ MVP rules are single-file context only for now
• ❌ MVP does not support autofixing for now

Demo

Here’s a demo of gptlint running on its own codebase:

Getting Started