v1.1.0: AI Agent Token Savings Focus

🎯 Major Refactor: Focus on AI Agent Token Savings

This release represents a complete refactor of context-rag to focus on its core value proposition: helping AI agents save 90% of tokens by providing relevant project context instead of sending entire codebases.

✨ What's New

🎯 Core Value Focused

• 90% Token Savings: AI agents get only relevant context (200 tokens vs 2000+)
• Project-Specific Responses: AI gets actual project patterns instead of generic answers
• Universal Compatibility: Works with OpenAI, Claude, Gemini, and any AI service

🔧 Simplified CLI

• context-rag ai "question" - Get project context for AI agents (token-efficient JSON)
• context-rag query "question" - Manual search with human-friendly output
• context-rag init - Zero-config setup
• context-rag index - Build semantic index
• context-rag status - Check index status

🧹 Cleanup & Simplification

• Removed over-engineered features (expanded search, complex plugins)
• Simplified documentation to focus on core workflow
• Clean, focused codebase
• Clear value proposition in README

🚀 Quick Start

# Install
npm install -g context-rag

# Setup (one-time)
cd your-project
context-rag init
context-rag index

# AI agents use this
context-rag ai "how to implement authentication"
# → Returns focused project context for massive token savings

💡 How It Works

Traditional Approach (Wasteful):

• User question (10 tokens) + entire conversation history (2000+ tokens) = 2010+ tokens
• Generic response that may not match project

With Context-RAG (Efficient):

• User question (10 tokens) + relevant context (200 tokens) = 210 tokens
• 90% token savings + project-specific response

📚 Documentation

• AI Agent Integration - How AI agents use context-rag for token savings
• API Reference - Complete command documentation
• Configuration - Setup and options

🎯 Target Users

• Developers using AI tools (ChatGPT, Claude, Gemini)
• AI agent developers looking to reduce token costs
• Teams wanting project-specific AI responses
• Anyone tired of sending entire codebases to AI services

---

This version is published to npm as context-rag@1.1.0 📦