Skip to main content
Learn how to choose the correct streaming wire format for your product and modality. Both Ambient/Form filling and Dictation use JSON text frames only (one JSON object per send).
Do not send raw binary audio frames.
To choose between Ambient and Dictation, refer to Ambient vs Dictation streaming guide.

Ambient/Form filling Wire Format

Outbound GET /ws/stream messages: START_TIME, AUDIO with data, EVENT, and the RU9G end marker. Form filling uses the same Ambient wire format on this socket.

Dictation Wire Format

Outbound GET /ws/transcribe messages: AUDIO with audioData, then EVENT with AUDIO_END. No START_TIME and no RU9G.
To learn about what’s common and different between Ambient/Form filling and Dictation streaming architecture, refer to Streaming architecture for more information.

​
Prerequisites

  • A created Ambient, Form filling, or Dictation session for the endpoint you will use.
  • A valid Suki Token. See Partner authentication.

Stream Ambient Audio

How Ambient streaming works, JSON text rules, and when to use /ws/stream.

Stream Dictation Audio

How Dictation streaming works, prerequisites, and when to use /ws/transcribe.

Audio Capture Best Practices

Capture format, environment, gain, and network readiness before you stream.

Dictation Transcript Frames

Inbound partial and final transcript frames for /ws/transcribe.

Complete the Session

Close the socket, end with REST, and troubleshoot common errors.

​
Next steps

Ambient streaming wire format - /ws/stream message shapes and chunking. Dictation streaming wire format - /ws/transcribe message shapes and chunking. Complete the session after streaming - Close the socket, end the session, and troubleshoot errors.
Last modified on October 16, 2026