Skip to content

Latest commit

 

History

History
53 lines (36 loc) · 2.07 KB

README.md

File metadata and controls

53 lines (36 loc) · 2.07 KB
Raw

Cursor Chat Reader

Read and analyze VSCode CursorAI chat backups.

Overview

This tool helps you preserve and analyze your conversations with CursorAI. It processes chat history from VSCode's database backups, extracts meaningful content, and saves it in a readable markdown format with added analytics.

How it works

  1. The tool reads SQLite database files from CursorAI backups
  2. Analyzes conversations using natural language processing:
    • Identifies programming languages and technologies mentioned
    • Categorizes discussions (debugging, architecture, learning, etc.)
    • Extracts frequent terms while filtering out common words
    • Counts code blocks and other content statistics
  3. Saves each conversation as a markdown file with metadata and tags

Analysis Components

  • Content Analysis One: Tracks code blocks, lists, and text volume in conversations
  • Context Analysis Two: Detects technologies and conversation categories
  • Term Extraction: Identifies important terms while filtering out common words using stop-words lists
  • Stop Words: Maintains lists of common technical terms and words in both English and Russian that should be excluded from analysis to focus on meaningful content

Setup

Before running the script, copy all state.vscdb.backup files from your Cursor backup directory to the backup folder:

  • Windows: %AppData%\Cursor\Backups
  • macOS: ~/Library/Application Support/Cursor/Backups
  • Linux: ~/.config/Cursor/Backups

When copying multiple backup files, allow your system to automatically rename duplicates (e.g., state.vscdb.backup, state.vscdb.backup (1), etc.). This ensures all chat history is preserved.

Usage

npm install
npm start

Configuration

You can configure the script by editing the chatreader.config.js file:

  • Adjust minimum text length for saving conversations
  • Modify maximum number of terms to extract
  • Customize file naming templates
  • Configure directories for backups and output

License

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