Zipf

A Next.js web app that analyzes the assistant side of a shared ChatGPT conversation and checks how closely its word frequencies follow Zipf's law.

You can check out the deployment here.

Features

• Validate and normalize ChatGPT share URLs before processing

• Fetch assistant-only transcript segments from the conversation

• Clean and finalize the corpus for reliable Zipf word-frequency analysis

• Compute rank-frequency results and present them in an interactive results view

• Show the captured transcript so you can inspect exactly what was analyzed

Usage

Working on a Single Conversation

1. Open the app in your browser

2. Paste a ChatGPT share link into the conversation input

3. Click the analyze button to run the pipeline:

   • URL validation and normalization

   • Assistant transcript retrieval

   • Corpus cleanup and Zipf analysis

Working with Analysis Output

1. Review the generated Zipf metrics and rank-frequency breakdown

2. Expand the transcript panel when you want to inspect source text

   • Confirm only assistant segments were included

   • Verify segment boundaries and extracted text

   • Re-run with another share URL to compare behavior

Contributing

I appreciate all contributions, feel free to submit a pull request.

License

This project is licensed under the MIT License, see the LICENSE file for details.