This repository provides tools and curated archives for analyzing the evolution of ChatGPT's frontend codebase, specifically its webpack chunks. It's designed for researchers, developers, and curious users interested in understanding how the application changes over time, offering a historical record and scripts for easier diffing and analysis.

How It Works

The project archives original, unmodified webpack chunk files in orig/ and provides unpacked, formatted versions in unpacked/ for easier comparison. It includes Node.js scripts to automate fetching new chunks, converting build manifests, unpacking and formatting files using Biome, and filtering URLs against a changelog. This approach facilitates detailed diffing of code changes between different ChatGPT builds.

Quick Start & Requirements

• Install dependencies: npm install (for Node.js scripts) and npx biome format --write unpacked/ (for formatting).
• Requires Node.js and git.
• Fetching new chunks involves manual steps using browser DevTools and piping output to provided scripts.

Highlighted Details

• Comprehensive CHANGELOG.md tracks build version changes.
• Scripts automate fetching, unpacking, and formatting of webpack chunks.
• Facilitates code diffing between builds using git diff on unpacked files.
• Includes links to related research on reverse engineering web applications.

Maintenance & Community

The project is maintained by 0xdevalias. Discussions can be found on Twitter, Reddit, and Hacker News via links in the README.

Licensing & Compatibility

The project is subject to multiple licenses with specific exceptions, detailed in LICENSE.md. Usage is intended for educational and research purposes, with a disclaimer against unauthorized or unlawful use.

Limitations & Caveats

The README notes that detailed documentation for helper scripts is not provided, requiring users to understand their operation. Manual intervention is required for fetching and saving new chunks. Biome formatting may require multiple passes.