NexusFlow.ai

NexusFlow.ai is a comprehensive agent orchestration platform that supports multiple AI frameworks (LangGraph, CrewAI, AutoGen, DSPy) through an adapter pattern.

Architecture

NexusFlow uses a modular architecture with framework adapters that allow you to build flows using any supported orchestration framework.

Features

• Multi-Framework Support: Build flows using LangGraph (fully supported in MVP), with basic support for CrewAI, AutoGen, and DSPy
• Visual Flow Designer: Drag-and-drop interface for creating agent workflows
• Flexible Tool Integration: Register, configure, and use custom tools within your flows
• Execution Monitoring: Track and visualize execution paths with real-time updates
• Flow Deployment: Deploy flows as API endpoints for production use

Tool Registry

NexusFlow includes a built-in tool registry that allows you to:

• Register custom tools that can be used across different frameworks
• Test tools individually before integrating them into flows
• Manage tool compatibility with different frameworks
• Configure framework-specific settings for each tool

Local Development Setup

Environment Variables

In addition to the database and API key settings, you can configure:

• DISABLE_AUTH=true - Disables authentication for local development
• USE_MOCK_TOOLS=true - Uses mock implementations for tools without API keys
• ALLOW_CODE_EXECUTION=false - Controls whether code execution tools are enabled
• REACT_APP_USE_MOCK_API=true - Enables mock API mode in the frontend

Prerequisites

• Python 3.9 or newer
• Node.js 16 or newer
• PostgreSQL 13 or newer

Backend Setup

1. Clone the repository:

   git clone https://github.com/yourusername/nexusflow.git
   cd nexusflow

2. Set up a Python virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Python dependencies:

   cd backend
   pip install -r requirements.txt

4. Configure your environment variables:

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

5. Initialize the database:

   alembic upgrade head

6. Run the backend development server:

   python run.py

   The API will be available at http://localhost:8000.

Frontend Setup

1. Navigate to the frontend directory:

   cd ../frontend

2. Configure your environment variables:

   cp .env.example .env
   # Edit .env if needed (REACT_APP_API_URL=http://localhost:8000/api)

3. Install dependencies:

   npm install

4. Start the development server:

   npm start

   The frontend will be available at http://localhost:3000.

Using Mock Mode

During development, you can use the mock API mode to work without a backend:

1. Set REACT_APP_USE_MOCK_API=true in your .env file
2. Or toggle mock mode in the login screen

Testing

Backend Tests

cd backend
pytest

Frontend Tests

cd frontend
npm test

Working with Frameworks

LangGraph

The primary framework supported in the MVP. To use LangGraph with NexusFlow:

1. Install LangGraph: pip install langgraph
2. Configure your OpenAI API key in .env
3. Create a new flow and select LangGraph as the framework
4. Add agents and tools to your flow
5. Configure agent roles and relationships
6. Test and deploy your flow

CrewAI, AutoGen, and DSPy

Additional frameworks will be fully supported in future releases. Basic integration is available in the MVP.

Deployment

Local Docker Deployment

1. Build and start the containers:

   docker-compose up -d

2. Access the application at http://localhost:3000

Production Deployment

For production environments, we recommend:

1. Setting up a proper PostgreSQL database
2. Using HTTPS for all communications
3. Implementing proper user authentication
4. Setting up monitoring and logging
5. Deploying behind a reverse proxy like Nginx

Contributing

Please see CONTRIBUTING.md for details on how to contribute to NexusFlow.

License

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