简体中文 | English

A privacy-focused tool to clean telemetry and tracking from Augment Code VSCode extension while preserving full functionality.

Aug Cleaner is a Python utility that patches JavaScript files to remove telemetry and tracking mechanisms from the Augment Code VSCode extension. This tool prioritizes user privacy by intercepting data collection while maintaining all AI features.

• Complete Privacy Protection: Blocks all telemetry and tracking API calls
• Session Anonymization: Generates random session IDs for each API call
• User-Agent Hiding: Removes user-agent information to prevent fingerprinting
• Zero Performance Impact: Direct source-level patching with no runtime overhead
• Surgical Precision: Only blocks tracking, preserves all AI functionality
• Simple Usage: One command to patch any JavaScript file

git clone https://github.com/gmh5225/aug_cleaner.git
cd aug_cleaner

# Patch Augment extension file directly
chmod u+w ~/.vscode/extensions/augment.vscode-augment-*/out/extension.js
python aug_cleaner.py ~/.vscode/extensions/augment.vscode-augment-*/out/extension.js

This will:

1. Create backup: Automatically creates extension_ori.js backup
2. Patch original: Modifies the original extension.js file directly
3. Enable privacy: Blocks all telemetry while preserving AI functionality

• Backup created: extension_ori.js (original file preserved)
• File patched: extension.js (modified with privacy protection)
• Safety checks: Warns if backup exists or file already patched

Aug Cleaner works by:

1. Locating Target Function: Finds the async callApi() method in JavaScript files
2. Injecting Privacy Code: Inserts a compact privacy protection patch
3. Blocking Telemetry: Intercepts all report-* and record-* API endpoints
4. Anonymizing Sessions: Generates random UUIDs for each API call
5. Hiding User Info: Clears user-agent strings

• report-feature-vector - Feature vector reporting
• report-error - Error reporting
• record-session-events - Session event recording
• record-request-events - Request event recording
• record-preference-sample - Preference sampling
• All other report-* and record-* endpoints

1. Telemetry Blocking

   • Intercepts all tracking API calls
   • Returns success responses to maintain functionality
   • Zero data leaves your machine

2. Session Randomization

   • Generates unique session ID for each API call
   • Prevents behavior correlation across sessions
   • Makes user tracking impossible

3. User-Agent Hiding

   • Removes system fingerprinting data
   • Prevents device identification
   • Maintains complete anonymity

• Python 3.6+
• No external dependencies required

The injected code is highly optimized and compressed:

// Intercept telemetry calls
if (typeof s === "string" && (s.startsWith("report-") || s.startsWith("record-"))) { 
  return { success: true }; 
}

// Generate random session ID
const chars = "0123456789abcdef"; 
let randSessionId = ""; 
for (let i = 0; i < 36; i++) { 
  randSessionId += i === 8 || i === 13 || i === 18 || i === 23 ? "-" : 
                   i === 14 ? "4" : 
                   i === 19 ? chars[8 + Math.floor(4 * Math.random())] : 
                   chars[Math.floor(16 * Math.random())]; 
}

// Apply privacy settings
this.sessionId = randSessionId;
this._userAgent = "";

• Surgical Precision: Targets the single API entry point
• 100% Coverage: All HTTP requests go through callApi()
• Minimal Code: ~15 lines vs 100+ lines of network interception
• Zero Performance Impact: No runtime overhead
• Maximum Compatibility: Doesn't interfere with other functionality

• Backup First: Always backup original files before patching
• Disable Auto-Update: Turn off auto-update for Augment Code extension in VSCode to prevent patch overwriting
• Extension Updates: Re-apply patch after manual extension updates
• Functionality Preserved: All AI features continue to work normally
• Privacy First: No telemetry data leaves your machine

1. Open VSCode Extensions panel (Ctrl+Shift+X)
2. Find the Augment Code extension
3. Click the gear icon on the right side of the extension
4. Select "Disable Auto Update"

Contributions are welcome! Please feel free to submit issues and pull requests.

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

• Built for privacy-conscious developers
• Inspired by the need for transparent AI tools
• Dedicated to user privacy and data sovereignty

Disclaimer: This tool is for educational and privacy protection purposes. Users are responsible for compliance with applicable terms of service and local laws.