| name | tts-synthesize |
| description | Use when a user explicitly asks to convert text into spoken audio, read text aloud, or reply with voice output. |
TTS Synthesize
Overview
Generate speech audio with DashScope TTS using qwen3-tts-flash.
When to Use
- User explicitly asks to convert text to audio.
- User asks for a spoken reply or for text to be read aloud.
Do not use this skill unless the user clearly wants audio output.
Inputs
- Text to synthesize
- Optional output path if the caller wants a specific filename
Configuration
OPENCLAW_DASHSCOPE_TTS_API_KEY
OPENCLAW_DASHSCOPE_API_KEY fallback
OPENCLAW_DASHSCOPE_TTS_BASE_URL optional override
OPENCLAW_DASHSCOPE_TTS_MODEL optional override, defaults to qwen3-tts-flash
OPENCLAW_DASHSCOPE_TTS_VOICE optional default voice
Use
- Collect the exact text that should be spoken.
- Run:
node synthesize.mjs --text "Hello world"
- or
node synthesize.mjs --text-file /path/to/text.txt
- Return the generated audio file path.
- Send or reference that audio file in the channel reply.
Notes
- Keep TTS opt-in only.
- The helper saves audio into persistent workspace-backed storage unless an explicit output path is supplied.