 |
 |
 |
 |
The execution engine supports asynchronous processing using job queues, while the agent enables users to create workflows using natural language prompts.
- Visual drag-and-drop workflow builder
- Modular and scalable backend design
- Extensible node registry architecture
- Asynchronous workflow execution
- Queue-based processing with BullMQ & Redis
- Workflow persistence using PostgreSQL
- AI-powered workflow building agent
- MQTT trigger nodes for IoT event-driven workflows
- AI and image processing nodes.
- Node.js
- Express.js
- TypeScript
- PostgreSQL
- Prisma ORM
- Redis
- BullMQ
- React
- Vite
- React Flow
The backend is organized around a modular execution engine:
- Node Registry for dynamic node discovery and registration.
- Workflow Engine responsible for dependency resolution and execution order.
- Queue System powered by BullMQ for asynchronous and scalable execution.
- Database Layer using Prisma ORM for workflow, project, and execution persistence.
- REST API exposing workflow management and execution endpoints.
- Node.js
- PostgreSQL
- Redis
git clone https://github.com/NInJozz/floo
cd floo/backend
npm install
cd ../frontend
npm installBackend
npm run devFrontend
npm run devThe complete technical documentation, design decisions, architecture, and implementation details are available in the companion repository:
- 📄 Project Documentation (LaTeX): https://github.com/NInJozz/floo-docs
This project was developed as a Computer Engineering Graduation Project and is intended for educational and portfolio purposes.