When you want an AI to read your code, opening files one by one and copy-pasting them is tedious. Gitingest removes that chore with a single URL edit. Change “hub” to “ingest” in a GitHub repository URL (github.com/... → gitingest.com/...), and it returns the repository’s directory structure and file contents bundled into one text digest. Paste that into ChatGPT or Claude, and you can ask questions with the whole codebase as context. It is an open-source project released in November 2024 by Coderamp Labs, based in France, and it is available not only as a web tool but also as a command-line tool (CLI), a Python package, and a browser extension.
Key Features
- Works with a single URL rewrite: Just visit
gitingest.com/owner/repoinstead ofgithub.com/owner/repo. No account and no installation are required ─ the text is generated right in the browser - Output formatted for LLMs: The directory tree and the contents of each file are formatted so a language model can read them easily. Statistics such as file structure, extract size, and token count are shown, so you can judge whether the digest fits within your context window before pasting it
- Also available as a CLI and Python package: Install it with
pip install gitingestand process local directories or remote URLs from the terminal. Output can be written to a file or piped to stdout for use with other commands. The Python API offers both synchronous and asynchronous ingestion and works from Jupyter notebooks - Private repository support: Supply a GitHub Personal Access Token (PAT) and private repositories can be ingested as well. According to the official site, the PAT is never stored in the backend, it is used once for cloning and then discarded from memory, and cloned repositories are deleted after processing
- Browser extensions and self-hosting: Extensions are available for Chrome, Firefox, and Edge so you can convert a repository in one click from a GitHub page. A Docker Compose setup for self-hosting is also published, which suits teams that want everything to stay inside their own environment
- Exclusions and fine-grained control: Files covered by
.gitignoreare skipped by default. CLI options let you include submodules or include gitignored files when you need them
Pricing
| Plan | Price | Highlights |
|---|---|---|
| Web tool (gitingest.com) | $0 | Conversion in the browser, no sign-up |
| CLI / Python package | $0 | Published as open source, runs locally |
| Self-hosting | $0 (you cover infrastructure costs) | Run it yourself with Docker Compose |
Pricing is as of August 2026. Neither the official site nor the GitHub repository lists any paid plan. Check the official site for the latest information.
Pros & Cons
✅ Pros
- Remarkably simple ─ a one-word URL change, with almost nothing to learn
- Token counts are shown up front, so you can tell whether the digest fits your AI’s context limit before pasting
- Multiple entry points (web, CLI, Python package, browser extension) cover everything from manual use to automation with one tool
- It is open source, so if you would rather not send internal code outside, you can switch to the
pipversion locally or self-host it - Private repositories are supported, and the site states clearly how tokens are handled
⚠️ Cons
- For large repositories the digest becomes huge and will not fit into an AI’s context window as-is; you need to narrow the scope with exclusion settings
- The output is a bundle of text, not a summary or an architectural explanation ─ interpretation is left to the AI
- If you run someone else’s private code through the hosted web service, check it against your organization’s policies (local execution or self-hosting is the alternative)
- The design is GitHub-centric, and behavior with other Git hosts varies by environment
Comparison with Similar Services
| Criteria | Gitingest | Repomix | code2prompt | Downloading a ZIP from GitHub |
|---|---|---|---|---|
| Main usage | URL rewrite / CLI / Python | CLI / web / MCP | CLI | Manual |
| Installation | Not needed (web) | Needed (not for the web version) | Needed | Not needed |
| Output format | Tree plus file contents as text | Text / XML / Markdown | Prompt shaped by a template | Raw files |
| Token count shown | Yes | Yes | Yes | No |
| Price | Free | Free | Free | Free |
All of them are open source and free, so the choice comes down to where you want to call the tool from. Gitingest fits when you want to stay in the browser, code2prompt when you want to craft templates in detail, and Repomix when you want to pick the output format.
Who Is It For
- Developers who want an AI to read open-source code and answer “where does this processing happen?”
- Anyone taking over an existing project who wants the AI to summarize the overall structure first
- People who read library sources to confirm usage but find jumping around GitHub tiresome
- Writers who need to grasp a repository’s structure while producing technical articles or documentation
- Teams that want an AI to draft a starting point for code reviews or design discussions
Summary
Gitingest compresses the everyday task of “handing a whole codebase to an AI” into a single URL edit. The functionality itself is simple, but that simplicity is the point: there is almost nothing to learn before you start. A comfortable order is to try a small open-source repository in the browser first, then move to the CLI with pip install gitingest once it becomes part of your routine. For internal code, choose local execution or self-hosting rather than the web version.