{
"type": "init",
"config": {
"language": "en",
"encoding": "pcm_s16le",
"sample_rate": 16000,
"operating_point": "enhanced",
"enable_partials": true
}
}"0000FF00000000FF0000000001010101010101010000"{
"type": "finalize"
}{
"type": "close"
}{
"type": "partial_transcript",
"transcript": "halo apa",
"start_time": 0,
"end_time": 0.8,
"confidence": 0.82
}{
"type": "final_transcript",
"transcript": "Halo, apa kabar?",
"start_time": 0,
"end_time": 1.4,
"confidence": 0.95
}{
"type": "utterance_end",
"last_word_end": 1.4
}{
"type": "error",
"code": "invalid_audio_type",
"message": "Unsupported audio encoding",
"status": 400
}Speechmatics Realtime v2
Real-time speech-to-text transcription with Speechmatics over WebSocket. Stream binary audio and receive partial and final transcripts with word timing, confidence, and optional diarization.
{
"type": "init",
"config": {
"language": "en",
"encoding": "pcm_s16le",
"sample_rate": 16000,
"operating_point": "enhanced",
"enable_partials": true
}
}"0000FF00000000FF0000000001010101010101010000"{
"type": "finalize"
}{
"type": "close"
}{
"type": "partial_transcript",
"transcript": "halo apa",
"start_time": 0,
"end_time": 0.8,
"confidence": 0.82
}{
"type": "final_transcript",
"transcript": "Halo, apa kabar?",
"start_time": 0,
"end_time": 1.4,
"confidence": 0.95
}{
"type": "utterance_end",
"last_word_end": 1.4
}{
"type": "error",
"code": "invalid_audio_type",
"message": "Unsupported audio encoding",
"status": 400
}API key issued by SLNG. Pass as Authorization: Bearer <token> in the WebSocket upgrade request headers.
GET
Target world part override. Auto-selected if not provided. Available world parts: ap, eu.
ap, euInitialize a Speechmatics Realtime session with recognition configuration.
Stream a raw binary audio frame to be transcribed.
Force the current utterance to end without closing the stream.
Signal end of audio stream and close the connection.
Interim transcription result (when partials are enabled).
Finalized transcription segment.
End-of-utterance signal driven by the silence trigger.
Indicates an error occurred during recognition.
Was this page helpful?