Nextflow

NextFlow is a visual AI workflow automation platform that enables users to design, execute, and manage AI-powered workflows through an intuitive drag-and-drop interface. It allows users to connect different processing nodes into a workflow, automate execution, and monitor results without writing code.

The platform uses a node-based editor where each node represents a specific task, such as collecting user inputs, processing images, interacting with AI models, or displaying the final output. Workflows are internally represented as Directed Acyclic Graphs (DAGs), ensuring that nodes execute in the correct dependency order while allowing independent tasks to run concurrently for improved performance.

NextFlow provides a complete workflow management experience, including secure authentication, workflow creation and management, automatic saving of changes, undo/redo functionality, JSON import/export, and detailed execution history. Every workflow execution is tracked with node-level status updates, execution duration, outputs, and error reporting, making workflows easy to debug and monitor.

The execution engine is powered by Trigger.dev, which handles long-running background tasks such as image processing and AI inference. Image manipulation tasks are performed using FFmpeg, while AI-powered text and multimodal generation are integrated through Google Gemini. Workflow data is stored using PostgreSQL with Prisma ORM, and user authentication is handled securely using Clerk.

Key Features

• Secure user authentication with Clerk
• Visual drag-and-drop workflow builder
• Custom workflow nodes:
• Request Inputs
• Crop Image
• Gemini 3.1 Pro
• Response
• Directed Acyclic Graph (DAG) based workflow execution
• Parallel execution of independent nodes
• Trigger.dev background task execution
• Google Gemini integration for AI-powered processing
• Image processing using FFmpeg
• Auto-save workflow changes
• Undo and Redo functionality
• JSON Import and Export
• Real-time execution status updates
• Workflow execution history with node-level logs
• Responsive and modern user interface built with Tailwind CSS

Tech Stack

Frontend

• Next.js 14 (App Router)
• TypeScript
• Tailwind CSS
• React Flow
• Zustand
• Lucide React

Backend

• Next.js API Routes
• Prisma ORM
• PostgreSQL (Neon)

Authentication

• Clerk

Workflow Execution

• Trigger.dev

AI

• Google Gemini

Media Processing

• FFmpeg
• Transloadit

Problem Solved

Building AI-powered workflows often requires manually switching between multiple tools, repeatedly uploading files, invoking AI models, and transferring outputs from one step to another. This process is time-consuming, difficult to manage, and not easily reusable.

NextFlow addresses this challenge by providing a unified visual workflow builder where users can define complete AI pipelines using interconnected nodes. The platform automatically manages execution order, dependency resolution, background processing, and workflow history, allowing users to create reusable AI workflows that can be executed reliably with a single click.