A working academy, seeded and repeatable.
The demo signs you in as K. Ramesh at Krishna Art Academy — 62 students, several hundred recorded works, ten calibration anchors, and a handful of students carrying genuine flags.
One of them is Aarav Krishnan: six works across three months, four dimensions climbing steadily, and value range flat since week seven. Nothing about that is scripted. The detector fits a line through the last 4 works and finds it the same way it would find it in your own academy.
Press anything you like. Reset puts it back exactly as it was.
Where to go once you are in
- 1The cohortAll 62 students, sorted by who needs attention first, with the reason printed beside every number.
- 2Record a workUpload any drawing. It is measured in your browser, five scores are proposed with reasons, and you confirm or override each one.
- 3Aarav Krishnan’s trajectorySix works over three months. Four dimensions climb. Value range stops at week seven and the plateau fires.
- 4The exercise that answers itA brief aimed at the flagged dimension, editable before it is issued.
- 5The parent reportFirst work beside latest, what moved, what is next. The instructor approves it before it can be shared.
What is real and what is seeded
The students and their works are generated, because real student drawings are not ours to publish. Everything the product does with them is the real thing: the same scoring path, the same calibration anchors in the prompt, the same plateau arithmetic, the same confirm-and-override rules. If you upload one of your own drawings on the capture screen, it goes through exactly the pipeline these seeded works went through.