Introducing Flatty: Simplify Codebase for LLM Processing
For you Cursor warriors out there, I just published Flatty, a GitHub repo that flattens your codebase into a text file for easy LLM processing s you can put the whole thing into the context window of Claude or ChatGPT. Take a look and let me know what you think! The UX is pretty simple:
- 1.
Copy/paste one line of code into Terminal to install
- 2.
Navigate to the local Github repo or directory you want to flatten
- 3.
Type "flatty" and hit enter
The script outputs your codebase as a single text file like this into a folder called "flattened" on your local machine. If the output text exceeds 100k tokens, it breaks the repo up into smart chunks and outputs a series of text files. The system includes robust controls that let you fine-tune key parameters like maximum output size and chunking behavior, while providing real-time token calculations and estimates - giving you precise control if you want to tweak. @here https://github.com/mattmireles/Flatty
