Agentflow: Multi-Agent Virtual Office

A sophisticated multi-agent system for business automation and decision support, powered by LangGraph and advanced AI agents.

Key Features

• Specialized AI Agents: Cofounder, Manager, Sales, Marketing, Finance, Legal, and more
• Structured Communication: LangGraph-based state management for reliable agent coordination
• Advanced Memory Systems: Neo4j for private memory, Qdrant for vector/global memory
• Tool Integration: Built-in tools for web search, financial modeling, and legal compliance
• Self-Healing: Automatic error detection and recovery mechanisms
• Personality Profiles: Customizable agent behaviors and expertise areas

Architecture

Core Components

1. LangGraph Orchestrator

   • Manages agent workflows and state transitions
   • Implements error handling and quality control
   • Ensures structured communication between agents

2. Agent Types

   • Cofounder: Vision and strategy
   • Manager: Project coordination and task delegation
   • Sales: Customer acquisition and forecasting
   • Marketing: Content strategy and brand amplification
   • Finance: Financial modeling and ROI analysis
   • Legal: Compliance and document generation

3. Memory Systems

   • Neo4j: Private agent memory and knowledge graphs
   • Qdrant: Vector embeddings and semantic search
   • Redis: Caching and queue management

4. Tool Integration

   • Web search and content analysis
   • Financial modeling and forecasting
   • Legal document generation
   • SEO and marketing analytics

Getting Started

Prerequisites

• Python 3.9+
• Docker and Docker Compose (for local development)
• API keys for required services (OpenRouter, Qdrant, Neo4j, Redis)

Installation

1. Clone the repository:

   git clone https://github.com/aparnap2/agentflow.git
   cd agentflow

2. Set up the environment:

   cp .env.example .env
   # Edit .env with your API keys and configuration

3. Start the services:

   docker-compose up -d

4. Install Python dependencies:

   pip install -r requirements.txt

5. Initialize the database:

   python -m scripts.initialize_db

Agent Personalities

Each agent has a distinct personality and expertise area:

Agent	Personality	Key Focus	Temperature	Confidence Threshold
Cofounder	Visionary	Strategy	0.7	0.75
Manager	Organized	Coordination	0.5	0.8
Sales	Persuasive	Revenue	0.6	0.7
Marketing	Creative	Branding	0.7	0.6
Finance	Analytical	ROI	0.3	0.8
Legal	Precise	Compliance	0.3	0.85

Development

Running Tests

pytest tests/

Code Structure

agentflow/
├── backend/
│   ├── agents/           # Agent implementations
│   ├── tools/            # Custom tools
│   ├── workflows/        # LangGraph workflows
│   ├── memory/           # Memory management
│   └── main.py           # FastAPI application
├── frontend/             # Web interface
└── docker-compose.yml    # Development environment

Documentation

For detailed documentation, see:

• Agent Development Guide
• API Reference
• Deployment Guide

Contributing

Contributions are welcome! Please read our Contributing Guidelines for details.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

For questions or support, please open an issue or contact your-email@example.com