Haha, it’s not ready for wide inputs yet, the machine learning engine hasn’t even begun training, those inputs will have to be done manually by really chippers at some point. That’s and the 3D rendering engine are the most ambitious parts of the project.
The spot progression suggestion system is difficult because it needs be trained tens of thousands (maybe even more) of times by actual permutations of spots. It’s not as simple as just identifying what spot progression can be used with what color in some rules database, I’m attempting to see if we can give the engine “taste”, and we need machine learning for that.
anyway, I’m rambling.