OpenTypeless

Volcengine Doubao ASR

Doubao ASR is integrated through Volcengine OpenSpeech bigmodel streaming API with binary framed WebSocket protocol.

Official Documentation

Volcengine Bigmodel Streaming ASR API

Official API reference for bigmodel streaming speech recognition.

https://www.volcengine.com/docs/6561/1354869

Volcengine Bigmodel Streaming SDK Guide

Protocol and SDK integration notes for streaming ASR.

https://www.volcengine.com/docs/6561/1257584

Volcengine AK/SK Management

Official access key creation and management.

https://www.volcengine.com/docs/6291/65568

Get Credentials

  1. Create access credentials in Volcengine console (app key + access key as required by your ASR setup).
  2. Verify speech service activation and billing.
  3. Prepare credential in app_key:access_key format.

Configure in OpenTypeless

  1. Settings > Speech Recognition > Provider: Doubao ASR (Volcengine).
  2. Paste credential in app_key:access_key format.
  3. Run Test and verify latency is returned.

Recommended Configuration

OpenTypeless connects to OpenSpeech WebSocket and sends gzip-compressed framed audio packets.

Provider: volcengine-doubao
Protocol: WebSocket binary frame
Endpoint: wss://openspeech.bytedance.com/api/v3/sauc/bigmodel
Headers: X-Api-App-Key, X-Api-Access-Key, X-Api-Resource-Id
Resource ID: volc.bigasr.sauc.duration
Credential format in app: app_key:access_key

Important

⚠️
If key format errors appear, verify app_key:access_key exactly (single colon, no spaces).