A Rust CLI that packs a codebase into a single prompt for LLMs, with a source tree, Handlebars templates, and token estimates.
khromov/ai-digest
A CLI tool that bundles a codebase into a single Markdown file for uploading to Claude Projects or custom ChatGPTs.
- Stars
- 682
- Forks
- 57
- Open issues
- 4
- Last commit
- 23 days ago
What is it?
ai-digest walks a project directory and combines its files into one Markdown file, codebase.md by default. The output can then be uploaded as knowledge to a Claude Project or a Custom GPT so the model can answer questions about the code without copying files by hand.
By default it skips common build artifacts and configuration files. Projects can add .aidigestignore rules to exclude more, and .aidigestminify rules to keep a file listed in the output while replacing its contents with a placeholder, which helps with large generated files or data files. Whitespace removal can reduce token count, though it is disabled for whitespace-sensitive languages such as Python and YAML.
The tool also offers a file size report with bar charts, a watch mode that rebuilds the output when non-ignored files change, and a Node.js library API for generating digests programmatically. It is installed or run through npx.
Key features
- 01Combines all files in a directory and its subdirectories into one Markdown file.
- 02Ignores common build artifacts and configuration files by default, with custom ignore rules via .aidigestignore.
- 03Replaces file contents with placeholders for patterns defined in .aidigestminify.
- 04Removes excess whitespace to reduce token count.
- 05Shows file sizes with bar charts, optionally sorted by size.
- 06Watches files and regenerates the output on changes.
Use cases
- Uploading a project's source to a Claude Project so the model can answer questions about it.
- Building a custom ChatGPT whose knowledge base contains a snapshot of your code.
- Keeping a single up-to-date code snapshot for AI review by re-running the tool or using watch mode.
- Using it from Node.js code to generate digest content programmatically.
Quick start
- 1
Run the CLI in your project directory
npx ai-digest
In collections
Topics
FAQ
- What is ai-digest?
- A CLI tool that bundles a codebase into a single Markdown file for uploading to Claude Projects or custom ChatGPTs. ai-digest walks a project directory and combines its files into one Markdown file, codebase.md by default. The output can then be uploaded as knowledge to a Claude Project or a Custom GPT so the model can answer questions about the code without copying files by hand.
- Is ai-digest free to use? What license does it have?
- ai-digest is open source under the MIT license. Check the license terms before using it in commercial products.
- Is ai-digest still maintained?
- Actively maintained. The last commit to ai-digest was on Oct 15, 2026.
- What can I use ai-digest for?
- Uploading a project's source to a Claude Project so the model can answer questions about it. Building a custom ChatGPT whose knowledge base contains a snapshot of your code. Keeping a single up-to-date code snapshot for AI review by re-running the tool or using watch mode. Using it from Node.js code to generate digest content programmatically.
- How do I install ai-digest?
- Run the CLI in your project directory: npx ai-digest
- Where can I get ai-digest?
- The source code and installation instructions are on GitHub: https://github.com/khromov/ai-digest
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A CLI and web tool that packs an entire code repository into one AI-friendly file for feeding to LLMs like Claude or ChatGPT.
A React component library for rendering streamed LLM output, fixing broken markdown and supporting custom components and syntax-highlighted code blocks.
Turns a Git repository or local directory into a single prompt-friendly text digest for feeding code to LLMs.
A self-hosted MCP server that lets ChatGPT, Claude and other MCP hosts read, edit and run code in your local projects.
A CLI that scans your codebase and generates compact AI context files (CLAUDE.md, .cursorrules, AGENTS.md) so coding assistants spend fewer tokens exploring the project.