Record your voice, see the waveform, then verify against a configurable API (or simulate locally). Supports optional streaming via WebSocket.