🔍 MyCursorData

Extract and export your Cursor IDE conversation history.

MyCursorData reads Cursor's internal SQLite database and exports all your AI conversations to readable text files, organized by project.

✨ Features

• 🔍 Auto-discovery - Automatically finds your Cursor databases on macOS, Linux, and Windows
• 📁 Project organized - Exports conversations grouped by workspace/project
• 🛠️ Tool call tracking - Includes full tool call details (file edits, terminal commands, searches)
• ✅ Approval tracking - Shows which changes were approved/rejected by the user
• 📊 Statistics - Get quick stats about your conversation history

📦 Installation

With pip

pip install mycursordata

With uv (recommended)

uv pip install mycursordata

Run directly with uvx

uvx mycursordata ~/my-cursor-export

🚀 Usage

Quick Export

Export all conversations to a directory:

# Export to a directory
mycursordata export ~/cursor-export

# Or simply (export is the default command)
mycursordata ~/cursor-export

Discover Databases

Find all available Cursor databases on your system:

mycursordata discover

Output:

🔍 Searching for Cursor databases...

Default search paths:
  ✅ /Users/you/Library/Application Support/Cursor/User/globalStorage/state.vscdb
  ❌ /Users/you/Library/Application Support/Cursor Nightly/User/globalStorage/state.vscdb
  ...

📊 Found 1 database(s):

  1. [GLOBAL] cursor/global
     Path: /Users/you/Library/Application Support/Cursor/User/globalStorage/state.vscdb
     Conversations: 42, Messages: 1337

View Statistics

Get quick stats about your conversation history:

mycursordata info

Export with Options

# Use a specific database file
mycursordata export ~/output -d /path/to/state.vscdb

# Show summary before exporting
mycursordata export ~/output --summary

# Verbose output for debugging
mycursordata export ~/output -v

📄 Output Format

Conversations are exported as text files organized by project:

output/
├── my-project/
│   ├── 2024-01-15_10-30-00_Add_dark_mode_toggle.txt
│   └── 2024-01-16_14-22-10_Fix_login_bug.txt
├── another-project/
│   └── 2024-01-17_09-00-00_Refactor_API.txt
└── unknown/
    └── 2024-01-18_11-45-00_abc123.txt

Each file contains:

• Conversation metadata (name, timestamps, message count)
• Full message history with thinking blocks
• Tool calls with arguments and results
• Code blocks and diffs
• Approval summary (what was approved/rejected)

🔧 How It Works

Cursor stores its conversation history in a SQLite database called state.vscdb. This tool:

1. Discovers the database by checking common locations for your OS
2. Parses the key-value store to extract conversation data
3. Groups conversations by project/workspace
4. Formats everything into readable text files

Database Locations

🐍 Python API

You can also use MyCursorData as a library:

from mycursordata import CursorDataParser, discover_cursor_databases

# Discover databases
databases = discover_cursor_databases()
for db in databases:
    print(f"Found: {db['path']} ({db['source']})")

# Parse and export
with CursorDataParser("/path/to/state.vscdb") as parser:
    # Get stats
    stats = parser.get_bubble_stats()
    print(f"Total conversations: {stats['conversations']}")
    
    # Export all conversations
    parser.export_all("./output")
    
    # Or get raw data
    conversations = parser.get_all_conversations()
    for project, convs in conversations.items():
        print(f"Project: {project}, Conversations: {len(convs)}")

📋 Requirements

• Python 3.11+
• No external dependencies (uses only stdlib!)

🤝 Contributing

Contributions are welcome! Please open an issue or PR on GitHub.

📜 License

MIT License - see LICENSE for details.