mcpField
A
aliyun-qwen-tts-realtime

aliyun-qwen-tts-realtime

github.com/cinienceCreative & media39659 installsUpdated 11w ago
Repository

About aliyun-qwen-tts-realtime

aliyun-qwen-tts-realtime is a skill for generating low-latency streaming speech with Qwen TTS models in Model Studio. It includes a demo to check realtime compatibility in the current SDK and runtime, with an optional non-realtime fallback.

Key features

  1. 01Generate streaming speech from text
  2. 02Select a supported realtime model
  3. 03Set voice and optional instruction
  4. 04Check SDK and runtime compatibility
  5. 05Use an optional non-realtime fallback

Use cases

  1. 01Produce low-latency spoken output
  2. 02Test realtime TTS in an environment
  3. 03Generate speech with a chosen voice
  4. 04Fall back when realtime calls are unsupported

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/cinience/alicloud-skills --skill aliyun-qwen-tts-realtime

After installing, type/aliyun-qwen-tts-realtimein Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

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

Details

From the project README

Category: provider

Model Studio Qwen TTS Realtime

Use realtime TTS models for low-latency streaming speech output.

Critical model names

Use one of these exact model strings:

  • qwen3-tts-flash-realtime
  • qwen3-tts-instruct-flash-realtime
  • qwen3-tts-instruct-flash-realtime-2026-01-22
  • qwen3-tts-vd-realtime-2026-01-15
  • qwen3-tts-vc-realtime-2026-01-15

Prerequisites

  • Install SDK in a virtual environment:
python3 -m venv .venv
. .venv/bin/activate
python -m pip install dashscope
  • Set DASHSCOPE_API_KEY in your environment, or add dashscope_api_key to ~/.alibabacloud/credentials.

Normalized interface (tts.realtime)

Request
  • text (string, required)
  • voice (string, required)
  • instruction (string, optional)
  • sample_rate (int, optional)
Response
  • audio_base64_pcm_chunks (array)
  • sample_rate (int)
  • finish_reason (string)

Operational guidance

  • Use websocket or streaming endpoint for realtime mode.
  • Keep each utterance short for lower latency.
  • For instruction models, keep instruction explicit and concise.
  • Some SDK/runtime combinations may reject realtime model calls over MultiModalConversation; use the probe script below to verify compatibility.

Local demo script

Use the probe script to verify realtime compatibility in your current SDK/runtime, and optionally fallback to a non-realtime model for immediate output:

.venv/bin/python skills/ai/audio/aliyun-qwen-tts-realtime/scripts/realtime_tts_demo.py \
  --text "This is a realtime speech demo." \
  --fallback \
  --output output/ai-audio-tts-realtime/audio/fallback-demo.wav

Strict mode (for CI / gating):

.venv/bin/python skills/ai/audio/aliyun-qwen-tts-realtime/scripts/realtime_tts_demo.py \
  --text "realtime health check" \
  --strict

Output location

  • Default output: output/ai-audio-tts-realtime/audio/
  • Override base dir with OUTPUT_DIR.

Validation

mkdir -p output/aliyun-qwen-tts-realtime
for f in skills/ai/audio/aliyun-qwen-tts-realtime/scripts/*.py; do
  python3 -m py_compile "$f"
done
echo "py_compile_ok" > output/aliyun-qwen-tts-realtime/validate.txt

Pass criteria: command exits 0 and output/aliyun-qwen-tts-realtime/validate.txt is generated.

Output And Evidence

  • Save artifacts, command outputs, and API response summaries under output/aliyun-qwen-tts-realtime/.
  • Include key parameters (region/resource id/time range) in evidence files for reproducibility.

Workflow

  1. Confirm user intent, region, identifiers, and whether the operation is read-only or mutating.
  2. Run one minimal read-only query first to verify connectivity and permissions.
  3. Execute the target operation with explicit parameters and bounded scope.
  4. Verify results and save output/evidence files.

References

  • references/sources.md

Comments

Markdown supported