Claude Long-Running Analysis Platform

A powerful platform for conducting sophisticated, long-running analysis tasks using Claude AI, featuring dynamic system prompts and persistent session tracking.

Key Features

• 🧠 Advanced iterative analysis powered by Claude AI
• 🔄 Configurable analysis iterations with fine-grained control
• 🎯 Dynamic system prompts that evolve during analysis
• ⚡️ Smart rate limiting with 10s cooldowns
• 📊 Real-time progress monitoring
• 💾 Automatic session persistence
• 📝 Rich Markdown output
• 🔍 Comprehensive analysis history

Technical Stack

• Backend: FastAPI for high-performance async API
• AI Engine: Claude API integration
• Frontend: React with real-time updates
• Storage: File-based session management
• Rate Limiting: 10-second interval throttling

Getting Started

Prerequisites

• Python 3.8+
• Node.js 14+
• Anthropic API key
• npm or yarn

Setup

1. Clone repository:

   git clone https://github.com/yourusername/claude-long.git
   cd claude-long

2. Set up environment:

   python -m venv venv
   source venv/bin/activate  # Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Install frontend:

   cd frontend
   npm install

4. Configure .env:

   ANTHROPIC_API_KEY=your_api_key_here
   

Launch

1. Backend:

   uvicorn main:app --reload --port 8000

2. Frontend:

   cd frontend
   npm start

Access at http://localhost:3000

Usage

Analysis Workflow

1. Start

   • Define objective
   • Set initial prompts (optional)
   • Launch analysis

2. Monitor

   • Track real-time progress
   • Review iteration logs
   • Follow reasoning process

3. Control

   • Adjust system prompts
   • Track prompt evolution
   • Manage iterations

4. Sessions

   • Automatic persistence
   • Retrieve by ID
   • Complete history

API Reference

• POST /start_analysis

  • Start new session
  • Parameters: task, system_prompt, iteration_count

• GET /analysis_status/{session_id}

  • Get session status and history

• POST /update_system_prompt

  • Update active prompt
  • Parameters: session_id, new_prompt

• POST /process_iteration/{session_id}

  • Trigger next iteration

Best Practices

• Respect 10-second cooldowns
• Start with broad prompts
• Document prompt changes
• Save important session IDs
• Export for documentation

Development

Contributing

1. Fork repository
2. Create feature branch
3. Submit pull request

License

MIT License - See LICENSE file

Support

Use GitHub Issue Tracker