⚡ Invite-only beta · elite performance analytics

Coaching-grade match intelligence
from the video you already shoot.

Shuttler reads rally structure, tempo, and a coaching takeaway from ordinary single-camera footage — and stays honest about what one camera can’t call. Where the read is confident it says so; where a phone physically can’t tell (a line call), it holds rather than guesses.

● confident call ● held — can’t tell honestly
See the analyzer

Reads the rallies.

Beyond tracking, Shuttler interprets rally structure. On footage where the geometric pipeline scores 0.00, a zero-shot vision-LLM reads rallies at 0.58–0.73 across two venues — the top of that range on a venue it had never seen.

0.00 → 0.73 rally-read recall — pipeline → VLM

See a sample Insight Report ›

Knows what it can’t see.

True intelligence means admitting uncertainty. Rather than hallucinate a number, Shuttler holds obscured moments for review — line calls and exact landing are physically unrecoverable from one camera, so we name them, we don’t guess.

⚠ Held for review abstains on what one camera can’t answer

Safe for juniors.

Every shared movement dashboard is face-blurred by default before it can be shared — and an unsure call is blurred, not shown. The shared report is stats only, no footage. Privacy-safe for academy and junior footage.

Invite-only beta. Videos you upload are stored on this project’s Google Cloud to build an open dataset — only upload footage you have the right to share. Not affiliated with any tournament or organization.