Tastegram is an AI powered camera that will decide whether the pictures you take are worth saving or not.

How does Tastegram work? Tastegram evaluates pictures based on a series of Machine Learning models we have trained. There are two categories in our model: “good”, and “bad”. The “good” category was populated with 9 GB of pictures from Aesthetically Approved Sources like museums and photography agencies; the “bad” category was populated with 9 GB of pictures from the most followed social media picture accounts. This data was used to train different models with different aesthetic tastes, from the standard model, which is generous and inclusive, to the Art mode, which only likes the best of the best, to the Influencer mode, which dislikes art but likes pictures of food and cute dogs. The Machine Learning model was created with CreateML with data found on the Internets.

What is the point of this app?

Aren’t you tired of figuring out whether the picture you took is good or bad? Let the AI do the work for you!

Can I save my pictures?

Only if the AI decides the picture is worth saving.

How was this made?

100% Swift, using David Okun’s Lumina framework for the camera.


Tastegram does not collect any data at all.

Available on the App Store