How I built an AI Video Clipper with Next.js, N8N, and Python (Microservices Architecture)

I spent the last 8 months building YouClip, an AI video repurposing SaaS that turns long YouTube videos into viral shorts using AI scoring and face tracking.

It handles real users in production.

The Architecture

The system uses a Microservices approach orchestrated by Docker Compose:

1. Platform: Next.js 16 (Frontend & Auth)
2. N8N: The brain. It manages the workflow logic.
3. YT-Downloader: A Python service to handle downloads.
4. FFmpeg-Builder: FastAPI service for auto-reframing (9:16) and burning subtitles.
5. Control-Plane: Node.js service for state management using Redis.

Why Microservices?

Video processing is heavy. By separating the Downloader and Renderer into different containers, I can scale them independently without crashing the main Next.js app.

The Outcome

I recently landed a full-time AI Engineering role, so I'm shifting my focus away from this SaaS.

Instead of letting the code gather dust, I packaged the Entire Source Code + Docker Setup as a Starter Kit.

If you want to build an Opus Pro clone or study this architecture, you can grab the code here:

👉 https://purchase.youclip.id/

It includes all 7 microservices, N8N workflows, and a setup guide.

Let me know if you have questions!