Swishly is an early-stage mobile product being built to use on-device computer vision to help athletes review shot outcomes, release mechanics, and training patterns—without wearable sensors.
Our prototype combines object detection, pose estimation, and temporal modeling. Early beta testing will validate accuracy across real court conditions.
Autonomous rim and ball detection. Classifies swishes, back-rim rolls, and airballs while charting the apex curve with 3D parabolic reconstruction.
Break down release point, elbow flare, knee dip timing, and landing balance frame-by-frame. Instant auditory cues tell you what to fix before your next rep.
Visualize where you're lethal and where you're cold. Auto-segments Corner 3s, Elbow Jumpers, and Paint floaters across weekly training logs.
Great shooters aren't lucky—they are repeatable. Swishly continuously captures every joint vector and flags kinetic chain leaks.
Measures ball trajectory at separation moment to ensure optimal basket entry volume.
Guarantees upper arm aligns parallel with hoop vertical axis to eliminate left-right misses.
Calculates energy flow from lower body coil directly into the upward shot sweep.
Detects whether you release at the peak of your jump or on the way down.
Prototype interface
Explore four planned operating modes in the current Swishly product concept.
Prototype UI — displayed session values are illustrative and are not product-performance claims.
Lean your iPhone against a water bottle, bleacher, or mini tripod. Swishly's auto-calibration detects the rim & court perspective in 2 seconds.
No wearable wristbands, no sensor clips, no special Bluetooth basketballs. Just you, your natural jumper, and pure computer vision.
Listen to live voice shot calling ("Swish! Release angle 47°") and review frame-by-frame joint biomechanics right on the bench.
Swishly is in active research and development. The interface below communicates the product direction; performance will be validated during the planned early beta.
Developing the core computer-vision pipeline for player, ball, hoop, pose, and shooting-action analysis.
Translating the research pipeline into a mobile prototype designed around simple court setup and useful post-session feedback.
Preparing an early beta to test reliability, usability, privacy, and coaching value with athletes and trainers on real courts.
Swishly is an early-stage AI sports technology startup based in Suzhou, China, building a privacy-first mobile basketball training product for global users.
