Automated AI Video Pipeline Build

I want to stand up an end-to-end pipeline that can take a YouTube or TikTok link, analyse the source, and then turn out a fresh entertainment video composed of fully animated scenes, new characters, revised narration, and updated visuals — all without me clicking through a dozen tools every time.

Here is what I need the finished system to do:
• Ingest a public link from YouTube or TikTok and pull the media + metadata.
• Break the original down into logical beats (script, scene boundaries, key moments, soundtrack cues).
• Regenerate the script in a new entertainment-oriented style.
• Produce animated scenes that map to the revised script, choosing or creating assets automatically (e.g. via text-to-image/video models such as Stable Diffusion, Runway, Pika, etc.).
• Add AI voice-over with natural pacing, sync it, layer background music, and render at HD or higher.
• Push the final cut to a cloud bucket and return a shareable link or downloadable file.
• Run hands-off once triggered and scale out gracefully when several links come in at once.

Tech stack is flexible, but I expect to see modern orchestration (Python, Node, or Go micro-services, Docker/K8s, serverless, or similar), solid use of media libraries (ffmpeg, GStreamer), and programmatic access to generative models and TTS services. Robust logging, retries, and a clear README are important so I can deploy or hand off to another engineer later.

Deliverables (acceptance criteria)
1. Source-controlled codebase with README and setup scripts.
2. Infrastructure as code or container definitions ready to spin up in my AWS or GCP account.
3. Demo showing a YouTube and a TikTok link processed into fully animated entertainment videos with no manual edits between trigger and final render.
4. Brief hand-off call or video walkthrough.

If you have already wired together similar automation or toolchains, let’s talk specifics and timelines.