AI Agent System

A robust system of AI agents that collaborate to plan and develop software projects.

Architecture

Components

1. Planner Agent

   • Brainstorms business ideas
   • Creates detailed technical roadmaps
   • Handles project planning and specifications
   • Maintains conversation history
   • Provides structured analysis and recommendations

2. Developer Agent

   • Analyzes technical requirements
   • Creates implementation plans
   • Generates code with documentation
   • Performs code reviews
   • Maintains development conversation history

3. Main System

   • FastAPI backend for agent orchestration
   • Async task processing
   • Project status tracking
   • Error handling and logging
   • Chat history management

Setup

1. Install dependencies:

pip install -r AI_Agent_System/backend/requirements.txt

2. Set up environment variables in config/.env:

OPENROUTER_API_KEY=your_api_key_here

Local Testing

Run the agent interaction test script:

python AI_Agent_System/tests/test_agents.py

This will:

• Generate business ideas
• Create development plans
• Analyze requirements
• Generate sample code
• Perform code review
• Save chat histories for analysis

Chat histories are saved in AI_Agent_System/logs/ with timestamps for review.

API Endpoints

Project Management

• GET / - Health check
• POST /projects/ - Start a new project
• GET /projects/{project_id} - Get project status
• POST /projects/{project_id}/feedback - Provide feedback
• GET /projects/{project_id}/chat_history - Get agent chat histories

Request Examples

Start a new project:

POST /projects/
{
    "project_type": "web_application",
    "requirements": "Create a task management system",
    "constraints": {
        "technology": "Python/FastAPI",
        "timeline": "2 weeks"
    }
}

Provide feedback:

POST /projects/{project_id}/feedback
{
    "feedback": "Add authentication system and user roles"
}

Features

1. Conversation Memory

   • Both agents maintain chat histories
   • Context-aware responses
   • Traceable decision-making

2. Structured Output

   • Detailed analysis and plans
   • Code generation with documentation
   • Comprehensive code reviews

3. Error Handling

   • Robust error catching and logging
   • Status tracking for each project stage
   • Detailed error reporting

4. Async Processing

   • Background task execution
   • Non-blocking API endpoints
   • Progress tracking

5. Project Refinement

   • Feedback incorporation
   • Plan iteration
   • Continuous improvement

Development Workflow

1. Planner Agent:

   • Generates business ideas
   • Creates technical specifications
   • Refines plans based on feedback

2. Developer Agent:

   • Analyzes requirements
   • Creates implementation plans
   • Generates and reviews code

3. System:

   • Orchestrates agent interactions
   • Tracks project progress
   • Manages chat histories
   • Handles errors and logging

Best Practices

1. Testing

   • Run local tests before deployment
   • Review chat histories for quality
   • Verify agent interactions

2. Monitoring

   • Check system logs regularly
   • Monitor project statuses
   • Review chat histories

3. Maintenance

   • Keep dependencies updated
   • Review and clean logs
   • Monitor API performance

Future Improvements

1. Integration with version control
2. Enhanced error recovery
3. More specialized agent roles
4. Advanced project analytics
5. Real-time progress updates