Hooper

Hooper is an AI basketball analytics app from Built Different AI Inc. (San Francisco, founded 2024) built around a single-phone record → auto-analyze → auto-highlight workflow, with no wearables, multi-camera rigs, or court markers required. Its proprietary Sports Vision computer vision model identifies players and shot attempts, giving pickup players technical stats once reserved for professionals. Founders Steven Kang (Yale) and Mike Wu (Stanford PhD), backed by the a16z Speedrun 004 accelerator.

Core Capabilities

• Shot charts and stats: automatically outputs shot distribution, makes/misses, points, and field goal percentage; three-point percentage appears only when the arc can be marked
• AI highlights: automatically strips dead-ball time and cuts per-player highlights into vertical videos sized for Reels/TikTok/YouTube
• Player profiles and tracking: accumulates individual stats across games, with friend and group leaderboards and group highlights
• Teams and leagues: per-player stats and group chat; the Team plan adds the web-based Hooper Hub (full-game footage and per-possession data)
• Full-court tracking: two phones each film one half and merge into a complete full-court session (two accounts required)
• Hooper API: open to a limited set of partners, no public documentation, apply by email

Pricing Three freemium tiers: Free $0/month (2 hours of recording, 5 saved clips), Pro $11.99/month (10 hours, unlimited clips, priority processing), Team $29.99/month (10 hours shared, unlimited seats, web team dashboard included). Subscriptions run through App Store/Google Play in-app purchase. Note: third-party directory sites still circulate the old '$9.99/month' pricing — that is outdated information; defer to the official site.

Company Background Built Different AI Inc. was founded in San Francisco in 2024, joined the a16z Speedrun 004 accelerator, and officially reports 38% day-30 retention in beta (versus an 8% industry average); it has presented at the Sloan sports analytics conference and partnered with Jeremy Lin's JLIN NextGen Camp. The rumored $2.0M Seed round traces to a single source and remains unverified.

Risk Notes Officially claimed accuracy is about 92% under the recommended setup, but results vary with lighting, camera height, and occlusion; a floor-level camera loses the shot chart and the two-/three-point distinction. Full-court requires two phones, which weakens the single-phone selling point. App Store US has only 5 ratings and Google Play roughly 3k installs — a small public feedback base for an early-stage product, and the data is not suitable as official game records. The app's privacy collection is broad (location, contacts, photos/videos, and more), so privacy-sensitive users should take note.