abinthomasonline/repo2txt

What it solves

It simplifies the process of providing entire codebases to Large Language Models (LLMs) by converting software repositories into a single, plain-text format suitable for prompts. This eliminates the manual effort of copying and pasting multiple files when seeking AI-assisted development help.

How it works

The tool runs entirely in the browser to ensure privacy. It connects to repositories via GitHub (including private ones with tokens), GitLab, Azure DevOps, or by uploading local directories and zip files. Users can filter files by extension, respect .gitignore patterns, or manually select folders. The tool then aggregates the selected code, calculates token counts using gpt-tokenizer, and generates a text output that can be copied or downloaded.

Who it’s for

It is designed for AI developers and software engineers who use LLMs to analyze, refactor, or document existing codebases.

Highlights

• Privacy-First: 100% browser-based processing; code never leaves the user's device.
• Diverse Sources: Supports GitHub, GitLab, Azure DevOps, local folders, and zip uploads.
• Smart Filtering: Includes .gitignore support, extension filtering, and a visual file tree for cherry-picking files.
• Performance: Uses virtual scrolling and web workers to handle repositories with over 10,000 files efficiently.
• LLM Ready: Features a real-time GPT token counter to help users stay within model context limits.