I was wondering if an API that produced such images would be useful to people who sell physical products, but who do not want to go through the hassle of manually creating these "contextual" images (eg, hire a person to hold or stand next to the product for a photograph, or pay for a photoshopped image).
Nearly all manufacturers, retailers, _sellers_ have the canonical white-background image of the product, and they could boost their sales, reduce returns, and provide an overall better user experience by simply sending the canonical image to an API at a cost of, say, $1 each.
They could try it beforehand for free, which is fine, because it will contain a "SAMPLE" watermark across the image. If they don't like the results, they need not purchase the "clean" image.
The API could allow customization of what to show for context (silhouette of person? a dog? a bike? your own thing? (just train it by moving by superimposing a few sample objects for us).
Seems like it checks all the boxes:
1) Big customer base. Anybody who sells stuff online and uses pictures.
2) Great UX. Anybody can drag said image into some UI and see a sample result, and then pay for the "clean" download of it. Developers can use an API.
3) Very specific problem with very specific solution.
4) Insane margins. Fully automated.
The only trouble seems to be sales. But by offering free images (with "powered by ProductImageApi" watermark), and/or by hustling with sales (where salesperson can easily include a link to the images the prospect _could_ buy right now), it seems like an easy sell.
Thoughts?
