Repomix
Pack your repo into one AI-friendly file that LLMs like ChatGPT and Claude can read in full.
For developers feeding whole repos to an LLM, few free tools deliver this much control per keystroke. The .gitignore awareness, Secretlint filtering, per-file token counts, and Tree-sitter compression are the reasons to pick it over pasting files together by hand. It's a terminal-first tool, so if you never touch a CLI, budget time to learn it — otherwise it's an easy default.
Verified 6d ago · liveness 69/100 · cite: rightaichoice.com/tools/repomix
- Developers feeding a full codebase to ChatGPT, Claude, or Gemini for refactoring and review
- Teams automating AI-ready repo snapshots in CI/CD with GitHub Actions
- Engineers integrating whole-repo context into AI assistants via the MCP Server
- Authors packaging blogs, docs, and books into LLM-compatible datapacks for Q&A
- Non-developers who won't work from a command line or Node install
- Teams wanting a hosted cloud dashboard that packs repos for them
- Users who need built-in AI chat rather than an output file to paste elsewhere
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Repomix if you're a non-developer who prefers a GUI, or if you need built-in AI chat capabilities, real-time collaboration, or cloud-hosted processing without a local install.
No direct costs since Repomix is free and open-source, but you must pay for an LLM subscription (ChatGPT, Claude, Gemini, etc.) to use the packed output.
Repomix is completely free and open-source, so the only cost is the LLM subscription you already have. Compared to GUI-based tools or managed services that charge per use, Repomix offers no per-seat or per-packaging fees, making it ideal for cash-strapped developers and startups.
In short
Repomix — Pack your repo into one AI-friendly file that LLMs like ChatGPT and Claude can read in full. Best for Developers feeding a full codebase to ChatGPT, Claude, or Gemini for refactoring and review, Teams automating AI-ready repo snapshots in CI/CD with GitHub Actions, Engineers integrating whole-repo context into AI assistants via the MCP Server. Free to use.
What's new in Repomix
Checked 4 days agoAcross the latest 1 update: 1 feature update.
What people actually say about Repomix — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
35 mentions across 2 sources (Hacker News, Lemmy) · researched Jul 3, 2026.
Average across the 2 sources that answered — each source counts once, not each post.
- +Free and open source (MIT license), runs locally with no server dependency.
- +Respects .gitignore and includes Secretlint for basic security scanning.
- +Tree-sitter-based compression removes comments and empty lines while preserving structure.
- +Outputs XML, Markdown, or Plain text for compatibility with diverse LLMs.
- +Token counting per file and total helps manage context window limits.
- −Token reduction is minimal compared to newer tools like Scribe (98%).
- −CLI-only with no GUI; may intimidate less technical users.
- −No iterative or agentic context management for multi-turn sessions.
- −Lacks advanced filtering beyond glob patterns and .gitignore.
- −Token counter may not perfectly match all LLM tokenizers.
- • No paid tiers; only hidden cost is time spent on rough edges and limited token optimization.
Viability Score
How well maintained and how widely used is Repomix? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: September 2026
How we score →Key Features
- Pack an entire repository into a single XML, Markdown, JSON, or plain text file
- Respects .gitignore automatically for git-aware packing
- Secretlint integration detects credential-like files and excludes them from output
- Token counting per file and for the whole repository
- Tree-sitter code compression for roughly 70% token reduction
- Option to remove comments and empty lines from packed code
- Glob-based include and ignore patterns to select files
- Pack remote repositories via --remote (URLs, branches, tree paths, commit URLs)
- Generate a persistent repomix.config.json with repomix --init
- Include file summary and directory structure in output
- Show line numbers in packed output
- Use as an MCP Server for AI assistant integration
- Chrome extension for packing codebases from the browser
- Docker image for isolated or containerized runs
- GitHub Actions for automated packing in CI/CD workflows
About Repomix
Repomix is an open-source tool that packs an entire codebase into a single AI-friendly file, giving large language models like ChatGPT, Claude, and Gemini complete repository context without manual file stitching. It's built for developers who work with AI on code review, refactoring, planning, and documentation, and for teams that want that context generated reproducibly from the command line. In practice you run npx repomix@latest or install it globally via npm, yarn, bun, or Homebrew, and it writes a repomix-output.xml to your directory. From there you can pack all or part of a repo using glob include/ignore patterns, or pack remote repositories with the --remote flag — which supports shorthands, full URLs, branches, specific tree paths, and even commit URLs. Because the CLI uses your local git it can reach private repositories too. Five mechanics do the real work. It respects your .gitignore so generated and vendored files stay out; Secretlint scans for known credential formats and drops matching files from the output; token counts are reported per file and for the whole repository so you can see how close you are to a context limit; Tree-sitter-based code compression extracts signatures and structure while stripping implementation detail for roughly 70% token reduction; and comments or blank lines can be removed separately. Output can be XML (default), Markdown, JSON, or plain text, and a repomix.config.json persists your settings. Beyond the terminal it ships as a Docker image for isolated runs, a Chrome extension, an MCP Server for AI assistant integration, and GitHub Actions for CI/CD automation, plus a way to use it as a Node.js library and custom instructions to embed prompts directly in the packed output. Compare it against naive concatenation or GUI collectors: Repomix still gives you the security and token controls those skip, and it remains free.
Behind the Verdict
Pick Repomix when the job is 'give the model the whole thing.' That covers refactor planning, bug investigation, third-party library security checks, and doc generation from an existing codebase. It also fits an unusual use case the docs call out: authors packaging blogs, docs, or book content into LLM-compatible datapacks for AI-powered Q&A. If you already work in a terminal, the startup cost is essentially zero. Pass on it if no one on the team will run a command. There's no cloud dashboard doing the packing for you, and no built-in chat — it produces a file you then hand to ChatGPT or Claude yourself. Teams expecting real-time collaborative editing, or people who want zero local install, should look elsewhere. Against pasting files together or a generic GUI collector, the differentiator is what's automated around the packing: .gitignore enforcement, Secretlint credential detection, per-file and repo-wide token counts, and Tree-sitter compression that the docs measure at roughly 70% token reduction. That set is why we'd reach for it on a large repo where token limits actually bite. Where it bites: the compressed output is signatures and structure, so if a question hinges on implementation detail inside a function body, you may need the uncompressed pack. Secretlint matches known credential formats — it is a safety net, not an audit. Recipe: run once with defaults to see token counts, add include/ignore globs and compression, then keep the result in CI with GitHub Actions. Using it as an MCP Server or Node.js library matters for engineers wiring repo context into an assistant or app rather than pasting it into a chat window. The recent JSNation Open Source Awards 2025 nomination under 'Powered by AI' is community recognition, not a feature — treat it as a signal
Researching Repomix? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Repomix actually fits — and what changes day-one when you adopt it.
You're starting a refactor of a legacy JavaScript project. You run 'npx repomix' in the project folder, then paste the resulting repomix-output.xml into ChatGPT with a prompt asking for a refactoring plan.
Outcome: ChatGPT analyzes the entire codebase at once, providing a comprehensive refactoring roadmap with specific changes, which you then implement.
You want to automate code review in your CI pipeline. You set up a GitHub Action that runs Repomix on every push, generating a packed file that's sent to an AI reviewer like CodeRabbit.
Outcome: Every pull request gets AI feedback on the whole codebase, reducing review time and catching issues early.
You need to audit a third-party repository for security vulnerabilities. You run 'npx repomix --remote <repository-url>' and pass the output to an AI assistant with a security-focused prompt.
Outcome: The AI flags potential credential leaks and insecure patterns, thanks to Repomix's Secretlint integration and complete context.
Use Cases
- Feed your entire React project to Claude for a comprehensive code review in one message.
- Debug production issues by providing the full backend codebase to ChatGPT with error logs.
- Generate documentation by packing your repository and asking Gemini to create a README and API docs.
- Refactor legacy code by sending the whole codebase to an LLM for analysis and change suggestions.
- Automate code review in CI pipelines using GitHub Actions to generate packed files for AI analysis.
- Pack a remote GitHub repository to quickly audit a third-party's code for security issues.
Limitations
- No underlying AI model; it's a codebase packaging utility.
- Output format is limited to XML, Markdown, and plain text.
- Packing may not handle large or binary files efficiently.
as of 2026-08-31
Verification history
We have re-verified Repomix 7 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 7 verification passes.
Free to cite with attribution — this page re-verifies continuously.
Where the pricing makes sense
The company stage and team size where Repomix's pricing actually pencils out — and where peers do it cheaper.
Repomix is completely free and open-source, so the only cost is the LLM subscription you already have. Compared to GUI-based tools or managed services that charge per use, Repomix offers no per-seat or per-packaging fees, making it ideal for cash-strapped developers and startups.
Setup time & first value
How long it actually takes to get something useful out of Repomix — broken out by persona, not the marketing-page minute.
Most users get value in under 5 minutes: just run 'npx repomix' in your project directory and it instantly generates the packed file. For persistent config, 'repomix --init' creates a config file in seconds. Docker users can start with a one-liner.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Repomix”, and we withheld 5: 5 could not be judged, because “Repomix” is a single word that other videos use for other things. Showing the 1 we can prove is about Repomix.
Official links
Tools that pair well with Repomix
Common stack mates teams adopt alongside Repomix, with the specific reason each pairing earns its keep.
Claude Code
Claude Code is Anthropic's agentic coding assistant that plans, edits, and runs commands across your repo from the terminal, IDE, or browser.
Replit Agent
Replit Agent turns a plain-English prompt into a runnable full-stack app inside your browser, then deploys it to a live URL.
Pieces for Developers
A desktop memory layer that captures your focused app every 2 seconds and pipes that history into MCP-ready AI assistants.
Featured Head-to-Head Comparisons
Repomix vs Bito
For solo developers needing a quick, free way to compress a single repo for LLM context, Repomix is ideal. For engineering teams working across many repos with AI coding agents, Bito’s live knowledge graph and agent integrations justify its paid model, especially with new Slack/Jira management features from 2026.
Repomix vs Poolside Ai
For individual developers needing a free, open-source tool to compress codebases for ChatGPT or Claude, Repomix is the clear choice. Poolside AI serves enterprises in regulated sectors with costly, capable agents and on-prem models, but it's overkill and inaccessible for smaller teams.
Repomix vs Cognition Ai
Repomix is the right choice for solo developers or small teams who need a free, lightweight tool to feed their entire codebase to AI assistants like ChatGPT—ideal for one-shot code analysis or refactoring. Cognition AI is built for enterprise teams that need an autonomous AI engineer to handle complex, multi-step tasks like bug triage, legacy modernization, and cross-platform builds at scale, backed by a $10M guarantee. Pick Repomix if you just want to prep your repo for AI; pick Cognition if you want the AI to do the engineering work end-to-end.
Alternatives to Repomix
View allClaude Code
Claude Code is Anthropic's agentic coding assistant that plans, edits, and runs commands across your repo from the terminal, IDE, or browser.
Replit Agent
Replit Agent turns a plain-English prompt into a runnable full-stack app inside your browser, then deploys it to a live URL.
Pieces for Developers
A desktop memory layer that captures your focused app every 2 seconds and pipes that history into MCP-ready AI assistants.
Frequently Asked Questions
Categories
Best-of guides
Used Repomix? Help shape our editorial sentiment research.