Systems
LiveMeta + X creative

AI Video Pipeline

Next.js + Remotion app for generating Meta and X video ads. ElevenLabs for voice, Whisper for transcripts, FLUX 1.1 / Gemini 3 Pro for stills, ComfyUI on a local RTX 3090 for the heavy lifting. ~$0.34 per finished 30-second ad.

Video ads got expensive at exactly the moment we needed to run more of them. Local GPUs and the right pipeline make $0.34-per-video a real thing instead of a slide.
Open live property Related niche systems
System Card
categoryContent tools
statusLive
page/ai-video-pipeline
AI avatar summary

Why AI Video Pipeline exists.

What AI Video Pipeline replaced, and what it does for me now. The internal dashboards stay private, so this is the public version of it.

Here's a summary of this page by my AI Avatar

Public version

Script to finished ad

01

Script and offer angle

02

Voice, captions, and stills

03

Remotion render and local GPU path

04

Meta/X creative variants

The live app is auth-gated, so the public proof is the production flow and generated media once a safe sample is approved.
What it proves

Video creative becomes a repeatable production pipeline instead of a slow vendor bottleneck.

Problem

PVM/GDM need more ad creative than a manual video workflow can produce economically.

1

What it does

Generates Ken Burns text-video ads, voiceovers, captions, and final renders programmatically. Optional cloud fallback (Replicate) when the local GPU is busy; default is the homelab.