Congrats on the launch! This is definitely a really interesting approach to data collection and integration of vision models. Is there a way to collaborate on creating the datasets as usually a lot of training data is required. I actually have my own product, ezML, which is a platform for computer vision integration for apps. ezML would be perfect for you, we would be able to lift the stress of hosting, training, and scaling these vision models and let you guys focus on the more prioritized tasks. In addition we would be able to greatly extend your functionality -- I think there can be some great synergy between so let me know if you are interested and we can set up a meeting :) -- Anyway, even if you aren't interested I really like this approach to computer vision integration and I wish you the best of luck!