I’ve been working with Sora for a while, mostly to create short promotional videos for products. Friends in e-commerce kept asking me if I could help them generate videos for ads and social media. Sora was impressive, but everyone ran into the same problems very quickly:
The exported videos always include a watermark
To remove it and access higher resolution or longer durations, you need a $200/month subscription
For many individuals and small creators, the monthly cost doesn’t map to their actual usage
Some countries can’t use the official platform at all
As an indie developer, I kept wondering why something so powerful was effectively locked behind a fixed subscription, even for people who only need a few videos occasionally. The tool is amazing, but the economics aren’t always aligned.
So I started building an alternative access layer on top of the official Sora APIs. The idea was simple:
pay only for what you generate — no subscriptions, no recurring fees, and no watermark by default. Credits never expire, and you can generate 1080p videos up to 25 seconds whenever you need them.
To give some context on pricing:
OpenAI’s Sora API is billed at $0.10 USD/sec, so a 15-second clip costs around $1.50 USD. In my implementation, a full video currently costs about $0.37 USD per clip. For individual creators who just need a handful of videos to test ideas or run small campaigns, that kind of pay-per-clip model tends to match their usage much better than a flat $200/month subscription.
What began as a personal workaround quickly turned into something my friends relied on for actual campaigns. That’s when I realized the demand wasn’t niche. There’s a growing group of creators who don’t want to become full-time prompt engineers or commit to enterprise-level pricing just to test concepts.
One observation I’ve made along the way: discussions around AI video often become polarized. Some people focus on its flaws or potential misuse, others dismiss it outright. My experience has been different — once you actually work with these tools, you start to see both their limitations and their potential. Like most technologies, they’re neither inherently good nor bad; the outcomes depend heavily on how people choose to use them. I’m personally more interested in exploring where this leads than standing on the sidelines.
This project is my attempt to make Sora video generation accessible for tinkerers, small marketers, and anyone who wants to experiment without a monthly bill hanging over their head.
Happy to answer technical questions about infrastructure or cost management if anyone’s curious.
watree•2mo ago
The exported videos always include a watermark
To remove it and access higher resolution or longer durations, you need a $200/month subscription
For many individuals and small creators, the monthly cost doesn’t map to their actual usage
Some countries can’t use the official platform at all
As an indie developer, I kept wondering why something so powerful was effectively locked behind a fixed subscription, even for people who only need a few videos occasionally. The tool is amazing, but the economics aren’t always aligned.
So I started building an alternative access layer on top of the official Sora APIs. The idea was simple: pay only for what you generate — no subscriptions, no recurring fees, and no watermark by default. Credits never expire, and you can generate 1080p videos up to 25 seconds whenever you need them.
To give some context on pricing: OpenAI’s Sora API is billed at $0.10 USD/sec, so a 15-second clip costs around $1.50 USD. In my implementation, a full video currently costs about $0.37 USD per clip. For individual creators who just need a handful of videos to test ideas or run small campaigns, that kind of pay-per-clip model tends to match their usage much better than a flat $200/month subscription.
What began as a personal workaround quickly turned into something my friends relied on for actual campaigns. That’s when I realized the demand wasn’t niche. There’s a growing group of creators who don’t want to become full-time prompt engineers or commit to enterprise-level pricing just to test concepts.
One observation I’ve made along the way: discussions around AI video often become polarized. Some people focus on its flaws or potential misuse, others dismiss it outright. My experience has been different — once you actually work with these tools, you start to see both their limitations and their potential. Like most technologies, they’re neither inherently good nor bad; the outcomes depend heavily on how people choose to use them. I’m personally more interested in exploring where this leads than standing on the sidelines.
This project is my attempt to make Sora video generation accessible for tinkerers, small marketers, and anyone who wants to experiment without a monthly bill hanging over their head.
Happy to answer technical questions about infrastructure or cost management if anyone’s curious.