mcpField
A
alicloud-ai-audio-asr-realtime

alicloud-ai-audio-asr-realtime

github.com/cinienceCreative & media39695 installsUpdated 11w ago
Repository

About alicloud-ai-audio-asr-realtime

A skill for realtime speech recognition with Alibaba Cloud Model Studio Qwen ASR. It prepares streaming requests that return transcript fragments and indicate when a result is final.

Key features

  1. 01Prepare realtime ASR request templates
  2. 02Support streaming session parameters
  3. 03Return transcript fragments
  4. 04Mark final transcript results

Use cases

  1. 01Generate realtime subtitles
  2. 02Provide duplex voice-agent input
  3. 03Transcribe streaming audio in browser or terminal clients

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/cinience/alicloud-skills --skill alicloud-ai-audio-asr-realtime

After installing, type/alicloud-ai-audio-asr-realtimein Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

agents/
openai.yaml
references/
sources.md
scripts/
prepare_realtime_asr_request.py

Details

From the project README

Category: provider

Model Studio Qwen ASR Realtime

Validation

mkdir -p output/alicloud-ai-audio-asr-realtime
python -m py_compile skills/ai/audio/alicloud-ai-audio-asr-realtime/scripts/prepare_realtime_asr_request.py && echo "py_compile_ok" > output/alicloud-ai-audio-asr-realtime/validate.txt

Pass criteria: command exits 0 and output/alicloud-ai-audio-asr-realtime/validate.txt is generated.

Output And Evidence

  • Save session payloads and response samples under output/alicloud-ai-audio-asr-realtime/.

Critical model names

Use one of these exact model strings:

  • qwen3-asr-flash-realtime
  • qwen3-asr-flash-realtime-2026-02-10

Use cases

  • Realtime subtitles and captions
  • Voice-agent duplex input
  • Streaming speech-to-text in browser or terminal clients

Prerequisites

  • Set DASHSCOPE_API_KEY in your environment, or add dashscope_api_key to ~/.alibabacloud/credentials.
  • Realtime sessions generally require WebSocket or streaming session handling in the client.

Normalized interface (asr.realtime)

Request
  • model (string, optional): default qwen3-asr-flash-realtime
  • language_hints (array, optional)
  • format (string, optional): e.g. pcm, wav
  • sample_rate (int, optional): e.g. 16000
  • chunk_ms (int, optional): frame size in milliseconds
Response
  • text (string): recognized transcript fragment
  • is_final (bool): finalization marker
  • usage (object, optional)

Quick start

Generate a request template:

python skills/ai/audio/alicloud-ai-audio-asr-realtime/scripts/prepare_realtime_asr_request.py \
  --output output/alicloud-ai-audio-asr-realtime/request.json

Operational guidance

  • Prefer 16kHz mono PCM unless your client stack requires another format.
  • Keep chunks small enough for responsive partial results.
  • If you only have recorded files, use skills/ai/audio/alicloud-ai-audio-asr/ instead.

References

  • references/sources.md

Comments

Markdown supported