🚀 CrazyContext

The ultimate context builder for LLM applications

Multi-source context gathering (web, papers, GitHub) with optional semantic search integration.

---

✨ Features

• 🌐 Multi-Source Gathering - Web, academic papers, GitHub repos, local files
• 🔍 Flexible Search - Keyword, fuzzy, or semantic (with Deep-Searcher)
• 📚 GitHub Integration - Search repos, download code, extract context
• 📄 Format Flexibility - HTML, Markdown, PDF, plain text
• 💾 Smart Caching - Download once, query forever
• 🎯 Profile System - Pre-optimized for research, coding, legal, etc.
• 🤖 LLM-Ready - Direct output for any LLM (no framework lock-in)
• 🔌 Optional Semantic Search - Integrate with Deep-Searcher when needed

---

🚀 Quick Start

Installation

# Core installation
pip install crazycontext

# With optional features
pip install crazycontext[fuzzy]           # Fuzzy search
pip install crazycontext[pdf]             # PDF support
pip install crazycontext[github]          # GitHub integration
pip install crazycontext[deepsearcher]    # Semantic search
pip install crazycontext[all]             # Everything

Basic Usage

from crazycontext import CrazyContext

# Initialize
cc = CrazyContext()

# Gather context from multiple sources
context = cc.from_topic(
    "machine learning",
    sources=["web", "papers", "github"],
    max_results=20
)

# Format for LLM
prompt = context.build(format="llm_prompt")
print(prompt)

---

📖 Examples

Download Documentation

# Download and convert to Markdown
cc.from_urls([
    "https://docs.python.org/3/tutorial/",
    "https://pytorch.org/tutorials/"
])

# Search downloaded content
results = cc.search("neural networks", search_type="keyword")

Academic Papers

# Fetch papers from arXiv, Semantic Scholar
context = cc.from_topic(
    "transformer attention mechanisms",
    sources=["papers"],
    max_results=10
)

GitHub Repositories

# Search and download GitHub repos
context = cc.from_topic(
    "pytorch neural networks",
    sources=["github"],
    github_options={
        "language": "python",
        "min_stars": 500
    }
)

Semantic Search (Optional)

from crazycontext.bridges import CrazyContextLoader

# Initialize with Deep-Searcher integration
loader = CrazyContextLoader()

# Gather with CrazyContext, search with Deep-Searcher
loader.load_from_crazycontext(
    topic="AI",
    sources=["web", "papers", "github"]
)

# Semantic query (uses FREE Gemini embeddings!)
answer = loader.query("What is deep learning?")
print(answer['answer'])

---

🎯 Use Cases

• Research - Gather papers, web articles, and code examples
• Documentation - Download and search technical docs
• Coding Assistants - GitHub code examples + documentation
• Knowledge Bases - Multi-source context for RAG systems
• Content Creation - Research and fact-checking

---

📚 Documentation

🎯 Start Here

• Getting Started Guide - Installation, quick start, basic concepts
• Examples - Real-world use cases
• CLI Guide - Command-line interface

📖 Reference

• API Reference - Complete API documentation
• Best Practices - Guidelines and recommendations
• llm.txt - LLM-optimized reference for AI assistants

📋 Full Documentation

• Documentation Index - Complete documentation overview

🗺️ Architecture & Design

• Unification Plan - Architecture & design
• Quick Reference - API cheatsheet
• GitHub Integration - GitHub features
• Deep-Searcher Bridge - Semantic search

---

🔧 Development

# Clone repository
git clone https://github.com/crazycontext/crazycontext.git
cd crazycontext

# Install in development mode
pip install -e .[dev]

# Run tests
pytest

# Format code
black crazycontext/

---

🤝 Contributing

Contributions are welcome! Please read our Contributing Guide first.

---

📄 License

Apache 2.0 - see LICENSE file for details.

---

🌟 Features Comparison

Feature	CrazyContext Alone	+ Deep-Searcher Bridge
Multi-source gather	✅	✅
Keyword search	✅	✅
Fuzzy search	✅	✅
Semantic search	❌	✅
GitHub integration	✅	✅
Cost	Free	~$0.001-0.01/query
Setup	Simple	Moderate

---

💡 Why CrazyContext?

• Simple - Works out of the box, no complex setup
• Flexible - Use standalone or with semantic search
• Cost-effective - Free core, optional paid features
• No lock-in - Works with any LLM framework
• Production-ready - Battle-tested, well-documented

---

Built with ❤️ for the LLM community

Making context building crazy simple! 🎯