llm-token-tracker

# Changelog All notable changes to this project will be documented in this file. ## [2.4.0] - 2025-10-07 ### Added - 🎉 **Gemini API Support** - Full integration with Google's Gemini models - Added `GeminiWrapper` class for automatic token tracking - Support for Gemini 1.0, 1.5, and 2.0 series models - Proper usage metadata extraction from Gemini API responses - 💎 **Gemini 2.0 Models** - Support for latest experimental models - `gemini-2.0-flash-exp` (Free during preview) - `gemini-2.0-flash-thinking-exp-1219` (Free during preview) - 📊 **Gemini Pricing** - Comprehensive pricing data for all Gemini models - Gemini 1.5 Pro: $0.00125/$0.005 per 1K tokens (input/output) - Gemini 1.5 Flash: $0.000075/$0.0003 per 1K tokens - Gemini 1.5 Flash-8B: $0.0000375/$0.00015 per 1K tokens - Gemini 1.0 Pro and Vision models - 🔧 **Enhanced Provider Detection** - Automatic Gemini client wrapping - Detects 'gemini' or 'google' in client name - Seamless integration with existing tracker API - 💰 **Cost Comparison Updates** - Compare Gemini with OpenAI and Claude - Added Gemini Flash and Pro to comparison tools - Updated MCP server with Gemini support ### Changed - Updated `TokenUsage` interface to include 'gemini' as a valid provider - Enhanced `calculateCost` function with Gemini pricing logic - Updated MCP server tools to accept 'gemini' provider - Improved documentation with Gemini examples and pricing tables - Version bump to 2.4.0 ### Documentation - Added comprehensive Gemini API example (`examples/gemini-example.js`) - Updated README.md with Gemini support details - Added Gemini pricing table with all model variants - Updated package.json keywords to include 'gemini' and 'google-ai' - Added peer dependency for `@google/generative-ai` ## [2.3.4] - Previous release ### Features - USD to KRW exchange rate support - File-based persistence - MCP server with intuitive session tracking - Support for OpenAI and Claude APIs --- For more details, see the full commit history on GitHub.