💰 Token Cost Calculator

Real-time token counting & cost calculator for AI models with INR pricing. Track costs, set warnings, and manage your AI budget directly in VS Code.

🎯 What Makes This Different?

While other extensions count tokens, Token Cost Calculator focuses on cost estimation and budgeting:

• ✅ Cost Calculation - See costs in INR (₹) and USD ($) for all your models
• ✅ Budget Tracking - Set warning limits for tokens and costs
• ✅ Multi-Model Comparison - Compare costs across GPT, Claude, and Gemini simultaneously
• ✅ Full Control - Configure your own models and pricing (no hardcoded defaults)
• ✅ Sidebar Tracking - Real-time cost tracking in sidebar panel
• ✅ Cache Pricing - Support for all cache types (input, write, read)

✨ Key Features

💵 Cost Calculation

• Real-time cost calculation in Indian Rupees (INR) and USD
• Shows costs for all configured models at once
• Supports cache pricing (cached input, cache write, cache read)
• Configurable USD to INR exchange rate

📊 Token Counting

• GPT (OpenAI): ✅ EXACT - Uses same tokenizer as OpenAI
• Claude (Anthropic): ⚠️ APPROXIMATE - May overestimate by 10-15%
• Gemini (Google): ⚠️ APPROXIMATE - May vary by 5-10%

🎨 Visual Features

• Status Bar Display - Shows tokens and cost with model name
• Detailed Breakdown Panel - Beautiful card-based layout with hover effects
• Sidebar Tracking - Real-time token and cost tracking
• Warning System - Visual warnings when limits are exceeded

⚙️ Configuration

• Full User Control - Add your own models with custom pricing
• Default Model Selector - Quick dropdown to change default model
• Reference File - Pre-configured models ready to copy-paste
• Welcome Guide - Comprehensive setup instructions

🚀 Quick Start

1. Install the Extension

Search for "Token Cost Calculator" in VS Code Extensions marketplace

2. Add Your Models

1. Press Ctrl+, to open Settings
2. Search: tokenCalculator.customModels
3. Click "Edit in settings.json"
4. Add your models (see examples below)

3. Start Using

• Select text → See tokens and cost in status bar
• Click status bar → See detailed breakdown for all models
• Right-click status bar → Change default model or open reference

📋 Example Configuration

{
  "tokenCalculator.customModels": {
    "gpt-4-turbo": {
      "displayName": "GPT-4 Turbo",
      "tokenizerType": "gpt",
      "inputPrice": 0.01,
      "outputPrice": 0.03,
      "cachedInputPrice": 0.005
    },
    "claude-3-opus": {
      "displayName": "Claude 3 Opus",
      "tokenizerType": "claude",
      "inputPrice": 0.015,
      "outputPrice": 0.075,
      "cachedWritePrice": 0.02,
      "cachedReadPrice": 0.003
    }
  }
}

🎯 Use Cases

• Budget Planning - Estimate costs before sending prompts
• Cost Comparison - Compare costs across different models
• Budget Alerts - Set warnings to avoid overspending
• Project Tracking - Track token usage and costs in real-time

📖 Documentation

• Quick Start Guide - Get started in 5 minutes
• Setup Guide - Complete setup instructions
• Configuration Guide - Detailed configuration reference
• Reference Models - Pre-configured models ready to use

⚙️ Settings

🆚 Comparison with Other Extensions

Our Focus: Cost estimation and budgeting
Their Focus: Token counting and visualization

⚠️ Important Notes

Tokenizer Accuracy

• GPT: ✅ EXACT (uses same tokenizer as OpenAI)
• Claude: ⚠️ APPROXIMATE (10-15% overestimate - Anthropic's tokenizer is proprietary)
• Gemini: ⚠️ APPROXIMATE (5-10% variance - SentencePiece not publicly available)

No Default Models

This extension gives you full control - you must configure your own models. This ensures:

• You use only the models you need
• You set your own pricing
• No hardcoded defaults that may be outdated

🛠️ Commands

• Token Cost Calculator: Show Welcome Guide - Open setup guide
• Token Cost Calculator: Change Default Model - Select default model
• Token Cost Calculator: Open Reference Models File - Open reference file
• Token Cost Calculator: Calculate Token Cost for Selection - Analyze selected text
• Token Cost Calculator: Reset Token Session - Reset tracking

📸 Screenshots

Status Bar

Shows real-time token count and cost with model name:

💰 1,500 tokens | ₹45 (GPT-4 Turbo)

Detailed Breakdown Panel

Beautiful card-based layout showing costs for all models with cache pricing options.

Sidebar Tracking

Real-time tracking with warnings when limits are exceeded.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📝 License

MIT License - see LICENSE file for details

🙏 Acknowledgments

• Uses tiktoken for GPT token counting
• Built with VS Code Extension API

Made with 💰 for developers who care about AI costs