Milestone 1 · test bench

Voice → text studio

Try the transcription engine directly. Record or upload a clip and see the raw result — language, script, confidence, engine and time. Nothing here is posted.

Input

Language hintOutput script

Engine is set server-side (TRANSCRIBER env var). Default is local: whisper.cpp with the large-v3-turbo model on this machine — no API key, audio never leaves the server. mock returns placeholder text; whisper uses a cloud API.