Stealth Token Optimizer
Invisibly reduce LLM token usage while visibly demonstrating savings. ⚡
Transform your AI development workflow with intelligent prompt optimization. Stealth Token Optimizer seamlessly integrates into VS Code to reduce token consumption—saving you money while maintaining prompt quality.
✨ Features
🎯 Intelligent Optimization
- Semantic Duplicate Detection - Removes redundant sentences while preserving meaning
- Code-Aware Optimization - Handles code blocks differently than natural language
- Provider-Specific Rules - Tailored optimization for OpenAI, Anthropic, or custom providers
- Configurable Aggressiveness - Conservative (15-25%), Balanced (25-40%), or Aggressive (40-55%) savings
📊 Real-Time Metrics
- Savings Dashboard - See tokens saved, cost reduction, and ROI
- Analytics Dashboard - Track optimization effectiveness over time
- Provider Comparison - Compare savings across OpenAI and Anthropic
- Historical Tracking - 7-day and 30-day trend analysis
🔧 Easy Configuration
- VS Code Settings - All options configurable in VS Code preferences
- Real API Integration - Connect to OpenAI and Anthropic for accurate token counting
- Secure Credential Storage - API keys stored securely using VS Code secrets
- User-Friendly Setup - One-click configuration wizard
🛡️ Security-First
- No Data Collection by Default - Telemetry is opt-in and can be disabled
- Secure Credentials - API keys encrypted at rest using VS Code's secrets API
- Local Processing - Optimization runs locally; no prompts sent to unknown servers
- Transparent - See exactly what's being optimized
🚀 Quick Start
Installation
- Open VS Code Extensions Marketplace
- Search for "Stealth Token Optimizer"
- Click Install
First Run
- Press
Cmd+Shift+P (Mac) or Ctrl+Shift+P (Windows/Linux)
- Type "Stealth Optimizer: Configure API Keys"
- Select your provider (OpenAI, Anthropic, or both)
- Enter your API key
- Select your preferred model
Test It Out
Cmd+Shift+P → Stealth Optimizer: Simulate Optimization Run
Watch as your prompts get optimized with estimated token savings displayed!
📖 Usage
Commands Available
| Command |
Description |
| Toggle Optimization |
Turn optimizer on/off with a click |
| Open Savings Dashboard |
View real-time metrics and trends |
| Open Settings |
Configure optimization preferences |
| Configure API Keys |
Set up OpenAI/Anthropic integration |
| View Analytics |
See detailed usage statistics |
| Simulate Optimization |
Test with demo prompts |
Configuration Options
Access via Cmd+Shift+P → "Preferences: Open Settings" → Search for "Stealth Optimizer"
| Setting |
Options |
Default |
Impact |
| Optimization Level |
Conservative / Balanced / Aggressive |
Balanced |
Controls savings percentage |
| Provider |
OpenAI / Anthropic / Custom |
OpenAI |
Tailors optimization rules |
| Detect Code |
On / Off |
On |
Preserves code structure |
| Semantic Threshold |
0.5 - 1.0 |
0.90 |
How similar lines must be to merge |
| Enable Telemetry |
On / Off |
Off |
Send anonymous usage stats |
📊 How It Works
The Optimization Pipeline
- Input - Your prompt (any length)
- Analysis - Detect code blocks, identify duplicates
- Optimization - Remove semantic duplicates, compress whitespace
- Output - Optimized prompt with token count comparison
- Tracking - Record savings to dashboard
Typical Savings by Level
Conservative: 15-25% savings (minimal changes)
Balanced: 25-40% savings (recommended)
Aggressive: 40-55% savings (power users)
Example
Original Prompt:
Review this code for performance. Check the code for optimization issues.
Look for performance problems. Identify performance concerns.
Tokens: 35
Cost: $0.00035
Optimized Prompt:
Review this code for performance. Look for performance problems.
Tokens: 13
Cost: $0.00013
Saved: 63% ⚡
🔐 Security & Privacy
Your Data is Safe
- Local Processing - Optimization happens on your machine
- No Cloud Storage - Prompts are never stored or sent anywhere
- Secure API Keys - Credentials encrypted using VS Code's secrets vault
- Telemetry Opt-in - Analytics only sent if explicitly enabled
What Gets Sent (If Telemetry Enabled)
- Optimization event metadata (no prompt content)
- Token counts before/after
- Duration and success status
- Anonymous user ID
What Never Gets Sent
- Your actual prompts
- Your API keys
- Proprietary code or content
- Personal information
⚙️ Troubleshooting
Dashboard Not Showing Data
- Click "Open Savings Dashboard" and run a test optimization
- Check that "Enabled" is turned on in settings
API Key Errors
- Verify your API key is valid at:
- Try removing and re-adding the key
No Savings Shown
- Ensure your prompt has some redundant or semantic duplicate content
- Try the "Simulate Optimization Run" command for a demo
- Check the optimization level (Conservative might be too strict)
- Disable "Detect Code" if you're primarily optimizing code
- Reduce "Semantic Threshold" to be less strict (more aggressive)
💡 Tips for Best Results
Maximize Savings
- Use Balanced level for consistent 25-40% savings
- Enable Code Detection for mixed code/language prompts
- Pair with verbose, instructive prompts (they have more duplication)
Maintain Quality
- Test optimized prompts with your LLM first
- Use Conservative mode for critical instructions
- Monitor dashboard to see what's being removed
Track ROI
- Note token counts before/after
- Check "View Analytics" to see aggregate savings
- Calculate monthly cost savings based on usage
📈 Real-World Impact
For a typical developer:
- 100 prompts/day
- 500 tokens average
- 30% savings with Balanced mode
- OpenAI's GPT-4 pricing ($0.03/1K input tokens)
Monthly Savings:
- Tokens saved: 1.5M
- Cost saved: ~$45/month
- Annual savings: ~$540
For an agency with 10 developers:
- Annual savings: ~$5,400
- Compounds with more prompts and larger models
🎯 Roadmap
Q1 2026
- ✅ Real-time token counting via provider APIs
- ✅ Analytics dashboard
- ✅ Multi-provider support
Q2 2026
- Dashboard trends and comparisons
- Optimization suggestions
- Performance metrics
Q3+ 2026
- IDE integration for other editors
- Enterprise analytics
- Custom optimization rules
📞 Support
- Issues: Report on GitHub Issues
- Questions: Check the FAQ in documentation
- Feature Requests: Open a discussion on GitHub
📄 License
MIT License - See LICENSE file for details
🤝 Contributing
We welcome contributions! Please see CONTRIBUTING.md for guidelines.
⭐ Show Your Support
If Stealth Token Optimizer helps you save on LLM costs, please:
- Give us a ⭐ on GitHub
- Leave a review on VS Code Marketplace
- Share your savings story with us!
Made with ❤️ by Stealth Utilities
Reduce token usage. Increase productivity. Grow profitably.
- Stealth Optimizer: Open Savings Dashboard
- Stealth Optimizer: Simulate Optimization Run (demo)
