I built an AI UGC video generator using my spare 5090. Looking for feedback on the business model and go to market.

Hey IH,
I have been exploring AI video for a while and finally turned my spare RTX 5090 into a working prototype.

I built an AI UGC video generator that creates creator-style videos from scripts, images, or audio. It produces those “talking to camera” videos you normally have to record yourself.

Link: https://aivideomaker.ai/ai-ugc

A few details about the build:
I can already generate 3 minute UGC videos on a single GPU, but the processing time takes up to an hour. Because of that, I limited the current version to 30 seconds. I am now scaling the system to 4 parallel GPUs so I can shorten rendering time and eventually bring back longer videos.

The motivation is simple. Most AI UGC tools cost 80 to 110 dollars a month, which ends up being around 10 dollars per video. I felt there should be a more affordable way to produce these videos, especially for ad testing and content iteration.

I would love suggestions from the IH community:
1. Business model
Should I charge by subscription (starting at 9.99 dollars) plus credits, or run it as a credits only system?
Which model feels more natural for an AI video product?

2. Go to market
Any ideas for reaching creators, DTC brands, or performance marketers?
Should I focus on TikTok creators, indie ecom operators, ad agencies, or something else?

Open to any feedback. Happy to answer questions about the tech or the process.