Visionati

I did some experiments, and the API is realy powerfull. It delivers greate results, and you can compare the output of different models, to see which one better fits your use case.
I saw in some cases it even provides coordinates of "objects" on the image.

The documentation for how to send API requests in the case of images available on the web is brief but sufficient for developers with some experience.

Could you add more details to the documentation on how to upload images via POST requests? Esp., which headers to use. It would be very much appreciated.

For the target group (developers), this is a great deal, esp. for controlling cost when developing proof of concepts and minimal viable products. It can also help you start your own app/SaaS with having running costs kick in only once your monthly credits are exhausted.