/live
/test
/test/ws, inspect real-time speech recognition transcripts, or run batch audio processing via /process.
/pipeline
/raw
/processUploads robot 8ch S16LE 16 kHz PCM. Runs the full DSR pipeline with default config (no toggle overrides). Returns clean mono 16 kHz WAV.
file (Required)curl -X POST "http://localhost:8123/process" \ -F "file=@input_8ch_16k.pcm" \ --output processed_clean.wav
/ws
/test/ws
/raw/ws
/live/ws/{target_client_id}