Drop in a dance, yoga, karate or fitness clip, and optionally the teacher's clip too. Get the SVT audit, PASSIONIT process scores, the PRUTL soul and materialism matrix and a Navarasa profile, rolled into one Kalki AIDharma Index. Then watch a bot dance the improved version to the same music.
Browsers can't pull video or music straight from YouTube, so paste the link for the record and upload the same clip (your own export or a screen recording). TestMe reads the movement and the beat from that file, on your device. Nothing is uploaded anywhere.
The two clips are lined up by matching their movement patterns over time (dynamic time warping), so a student who starts later or slows down is still compared move for move. The teacher video plays on that same matched timeline. Camera position and framing differ between recordings, so level and space are compared relative to each dancer's own centre.
Overall Structural Index
Positive columns: higher is better. Shadow columns (negative soul, negative materialism): lower is better. Net = average of the two positives and the two inverted shadows.
Rasa is read from body dynamics only: force, flow, stillness, levels and timing. Facial abhinaya and eyes are not read yet, so subtle rasas such as Karuna and Bibhatsa are estimates.
Kalki AIDharma Index
The mean of four layers: .
How scores are made: TestMe tracks frame-to-frame movement (how much, where in the frame, how sharply it starts and stops) and finds the beat from the soundtrack. Each pillar is computed from those signals. Intent and narrative are read from movement patterns, so treat them as a coach's first look rather than a final verdict.