HyperTwist — API.md

API philosophy

HyperTwist should use APIs to support the native runtime, not to replace it. The system should remain primarily native-first and local-first where reasonable.

Main API families

Recognition API

frame ingestion
facelet/state detection
confidence reporting
corrective suggestions

Replay API

replay storage and export
solve timeline annotations
AI-commentary attachment

Coaching API

drill recommendations
progression updates
error clustering
personalized next-step guidance

Profile / progression API

user settings
training history
unlocked modules
performance metrics

Curriculum API

guided lesson packets
method ladders
hyper-specific modules
school or educational variants if later needed

Deployment posture

Keep local computation local where sensible. Do not server-centralize everything just because an HTTP interface can be drawn.

947 B Raw Blame History