fengmk2/repomix

Repomix solves a common problem when working with AI assistants: getting your entire codebase into a conversation. Instead of copying and pasting individual files or struggling to explain your project structure, Repomix packages everything into a single file that AI tools like Claude or ChatGPT can understand quickly and thoroughly. Here's how it works. You run a simple command in your project directory, and Repomix reads through all your code files, respects your .gitignore settings so it doesn't include unwanted files, and bundles everything into one organized file. That file is formatted in a way that makes it easy for AI to parse and understand, think of it like preparing your code for an AI conversation the same way you'd organize documents before giving them to a colleague. You can then paste this single file into ChatGPT, Claude, or any other AI tool along with your question, and the AI has full context about your codebase. The tool is genuinely simple to use. If you have Node.js installed, you can run it without even installing anything by typing npx repomix@latest in your project folder. Or you can use the website at repomix.com, a browser extension for GitHub repositories, or even a VSCode editor plugin if you prefer a graphical interface. You can customize what gets included or excluded, count how many tokens your codebase will use (important for staying within AI model limits), and even compress the output to make it smaller. People use this for practical tasks: asking an AI to review their code architecture, generate missing tests, write documentation, or help with refactoring. Developers also use it to analyze open-source projects they want to contribute to, or to quickly onboard the codebase into discussions with collaborators. The project includes built-in security checks to prevent accidentally including secrets or passwords in your output file, an important safeguard when sharing code with AI services.