A shadow coach that listens to how care actually lands, not just what was said. Providers record patient conversations with a paired Plaud NotePin, and the app turns each one into concrete coaching. Plaud transcribes the audio with speaker diarization; Interhuman AI analyzes the same audio for social signals — confusion, hesitation, agreement, disengagement — per 10-second window, plus a five-dimension Conversation Quality Index (clarity, authority, energy, rapport, learning). A Crusoe-hosted LLM then combines the transcript with those signals to generate specific, actionable coaching: not "speak more clearly," but "at 2:15 you explained the dosage, confusion signals appeared right after — try rephrasing and asking if it's clear." A Neo4j knowledge graph tracks each provider's scores and recurring coaching themes across sessions, so feedback compounds over time instead of resetting every visit.
CareCompass
An AI shadow coach for healthcare providers — turns every patient conversation into coaching insights on clarity, rapport, and engagement.

About
Interhuman API
POSThttps://api.interhuman.ai/v1/upload/analyze
docsAPIs / Products
Plaud Transcription APIInterhuman AI APICrusoe Inference APINeo4j AuraDB
Tech stack
SwiftSwiftUIFastAPIPython
Use case
HealthcareProvider coaching
Type
iOS app