Multilingual Text to Speech & Voice Cloning — 10 Languages

Turn text into natural speech in 10 languages, steer the delivery with a plain-English style note ("speak warmly and slowly"), or clone a voice from a short recording.

Good for: multilingual voiceovers, e-learning and course narration, ads and product videos, character voices, localisation of existing scripts.

Quick start

{
  "text": "Hello, welcome to the demo.",
  "language": "English",
  "voice": "Vivian",
  "instruct": "Speak warmly and clearly.",
  "format": "mp3"
}

Clone a voice instead:

{
  "text": "Now I read in the cloned voice.",
  "ref_audio_url": "https://example.com/my-voice.wav",
  "format": "mp3"
}

What you get

• 10 languages: English, Chinese, Japanese, Korean, German, French, Russian, Portuguese, Spanish, Italian
• Preset speakers — male and female, several styles
• Style control — describe the mood and pacing in your own words
• Voice cloning from a short reference clip
• Speed control, long-text support, MP3 / WAV / Opus output

Input

Field	Default	What it does
text	required	The text to read
language	Auto	Language of the text
voice	Vivian	Preset speaker (ignored when cloning)
instruct	—	Style note, e.g. "sound excited and energetic"
ref_audio_url	—	Link to the voice you want to clone
ref_text	—	Transcript of that clip — improves cloning quality
speed	1.0	Playback speed
format	wav	wav, mp3 or opus

Output

One dataset row per run: audio URL, duration, language, voice or clone, processing time.

Pricing

No rental fee — you only pay Apify compute.

Notes

• Vietnamese is not supported here — use the Vietnamese Text to Speech Actor.
• Only clone voices you own or have permission to use.

For AI agents (MCP-ready)

This actor is built to be called by AI agents. It works out of the box with the Apify MCP Server — add it to Claude Desktop, Cursor or any MCP client: the agent can produce multilingual speech and clone voices on its own.

{
  "mcpServers": {
    "qwen3-tts": {
      "url": "https://mcp.apify.com/?actors=andrew_babo/qwen3-tts",
      "headers": { "Authorization": "Bearer <YOUR_APIFY_TOKEN>" }
    }
  }
}

Agent skill (paste into your agent's instructions)

Use the "qwen3-tts" tool for text-to-speech in 10 languages (English,
Chinese, Japanese, Korean, German, French, Russian, Portuguese, Spanish,
Italian), with plain-language style control and zero-shot voice cloning.
Vietnamese is NOT supported — route it to "vieneu-tts".

HOW TO CALL
- { "text": "...", "language": "English", "voice": "Vivian",
    "instruct": "Speak warmly and clearly.", "format": "mp3" }
- Cloning: pass "ref_audio_url" plus "ref_text" (the exact transcript) for
  best quality.
- Keep "chunk_parallel": 1 on CPU; the model already uses all cores.

OUTPUT CONTRACT
- One row per run: audio_url, duration, language, voice or clone, timing.
- errorCode / noResults means empty input.
- Only clone voices the user owns or has permission to use.