explaingit

yamadashy/repomix

📈 Trending25,049TypeScriptAudience · developerComplexity · 2/5ActiveLicenseSetup · easy

TLDR

Command-line tool that bundles your entire codebase into a single AI-ready text file, with token counting and secret detection.

Mindmap

mindmap
  root((repo))
    What it does
      Bundles code into one file
      Counts tokens for AI limits
      Detects secrets before sharing
      Respects gitignore rules
    Input & Output
      Input: local repository
      Output: XML, Markdown, text
      Compression option available
      Browser extension support
    Use cases
      Ask AI to review code
      Get refactoring suggestions
      Explain existing codebase
      Add features with AI help
    Tech Stack
      TypeScript
      Node.js
      Command-line interface

Things people build with this

USE CASE 1

Ask an AI assistant to review your entire codebase and suggest improvements in one conversation.

USE CASE 2

Get refactoring recommendations by sending your project structure to Claude or ChatGPT without manual copy-paste.

USE CASE 3

Have an AI explain how your existing code works by providing the full context in a single formatted file.

USE CASE 4

Collaborate with an AI to add new features by giving it complete visibility into your project structure and dependencies.

Tech stack

TypeScriptNode.js

Getting it running

Difficulty · easy Time to first run · 5min
Use freely for any purpose including commercial, as long as you keep the copyright notice.

In plain English

Repomix is a command-line tool that bundles an entire code repository into a single text file formatted for AI assistants. When you want to ask an AI like Claude, ChatGPT, or Gemini to help you with a codebase, you often need to give it the relevant code as context. Repomix automates the tedious work of copying and pasting files by packaging everything into one AI-ready document. You run one command in your project directory and it outputs a single file containing all your code, respecting your .gitignore rules so it skips build artifacts and other irrelevant files. It counts the tokens for you, important because AI models have limits on how much text they can process at once. It can also compress the output using a code parser to reduce token count while preserving the important structural information. The tool outputs in XML, Markdown, or plain text format and includes a security check that scans for accidentally included secrets like API keys before you send the bundle to an external AI service. You would use Repomix when you want to ask an AI assistant to review, refactor, explain, or help add features to an existing codebase in one seamless conversation. It also has a browser extension for one-click use directly from GitHub repository pages. The tech stack is TypeScript, running on Node.js.

Copy-paste prompts

Prompt 1
I've bundled my codebase with Repomix. Here's the output: [paste file]. Can you review this code and suggest refactoring opportunities?
Prompt 2
Use Repomix to bundle my project, then ask an AI: 'Here's my full codebase. What are the main architectural patterns and potential issues?'
Prompt 3
Run repomix in my project directory to create an AI-ready bundle, then paste it to Claude with: 'Help me add a new authentication module to this codebase.'
Prompt 4
Generate a Repomix bundle of my repo and ask ChatGPT: 'Explain the data flow and dependencies in this project structure.'
Prompt 5
Bundle your code with Repomix, check the token count, then ask an AI to help optimize performance bottlenecks in the included files.
Open on GitHub → Explain another repo

Generated 2026-05-18 · Model: sonnet-4-6 · Verify against the repo before relying on details.