Qwen3-TTS — Local Inference (No Server)

Run Qwen3-TTS directly in Python — no HTTP server, no REST API. Call a script or import the engine in your own code.

Quick reference

Mode	What it does	Key args
custom-voice	9 built-in speakers, optional emotion/style	--speaker, --instruct
voice-design	Describe the voice in natural language	--instruct (required)
voice-clone	Clone from ~3 s reference audio	--ref-audio, --ref-text

Available Speakers

The CustomVoice model includes 9 premium voices:

Speaker	Language	Description
Vivian	Chinese	Bright, slightly edgy young female
Serena	Chinese	Warm, gentle young female
Uncle_Fu	Chinese	Seasoned male, low mellow timbre
Dylan	Chinese (Beijing)	Youthful Beijing male, clear
Eric	Chinese (Sichuan)	Lively Chengdu male, husky
Ryan	English	Dynamic male, rhythmic
Aiden	English	Sunny American male
Ono_Anna	Japanese	Playful female, light nimble
Sohee	Korean	Warm female, rich emotion

Languages: Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, Italian, Auto

---

1 — Setup

Install dependencies once (from the skill directory):

First-time setup (one-time):

bash scripts/setup.sh

Custom download location:

python scripts/download_models.py --model-dir /path/to/models

Models are stored under {baseDir}/models/ by default. Override with QWEN_TTS_MODEL_DIR env var or --model-dir flag.

---

2 — Generate speech (CLI)

Custom Voice (default)

cd {baseDir}
python scripts/tts.py "Hello, how are you today?" --speaker Ryan --language English

With emotion/style instruction:

python scripts/tts.py "Great news everyone!" --speaker Aiden --instruct "cheerful and energetic"

Voice Design

Describe the voice in natural language:

python scripts/tts.py "Welcome to our show!" \
  --mode voice-design \
  --language English \
  --instruct "Warm, confident female voice in her 30s with a slight British accent"

Voice Clone

Clone a voice from a short (~3 s) reference audio clip:

python scripts/tts.py "This is spoken in the cloned voice." \
  --mode voice-clone \
  --language English \
  --ref-audio path/to/reference.wav \
  --ref-text "Transcript of the reference audio."

Common options

Flag	Purpose
-o output.wav	Save to exact file path instead of auto-named file
--output-dir DIR	Override output directory (default: tts_output/)
--model-dir DIR	Override model directory
--json	Print result as JSON
-v	Verbose logging

---

3 — Python API

Use the engine directly in code:

import sys
sys.path.insert(0, "{baseDir}/scripts")

from inference import TTSInferenceEngine

engine = TTSInferenceEngine(
    model_dir="{baseDir}/models",   # optional, uses default if omitted
    output_dir="./tts_output",       # optional
)

result = engine.generate_custom_voice(
    text="Hello world!",
    language="English",
    speaker="Ryan",
    instruct="calm and professional",
)
print(result)
# {"file": "tts_output/custom_voice_20260218_...wav", "duration_s": 1.23, "inference_s": 4.56}

Available methods:

• engine.generate_custom_voice(text, language, speaker, instruct)
• engine.generate_voice_design(text, language, instruct)
• engine.generate_voice_clone(text, language, ref_audio, ref_text)
• engine.status() — returns loaded variant, device, paths

---

4 — Configuration

All settings are controlled via environment variables. Set them before running.

Variable	Default	Description
QWEN_TTS_MODEL_SIZE	small	small (0.6B) or large (1.7B)
QWEN_TTS_MODEL_DIR	{baseDir}/models	Where model weights are stored
QWEN_TTS_DEVICE	auto (cuda:0 or cpu)	Inference device
QWEN_TTS_DTYPE	auto (bfloat16 / float32)	Model precision
QWEN_TTS_OUTPUT_DIR	./tts_output	Where generated .wav files are saved

Switch to the 1.7B model:

set QWEN_TTS_MODEL_SIZE=large
python scripts/tts.py "Hello world"

Use a custom model directory:

set QWEN_TTS_MODEL_DIR=D:\my-models\qwen-tts
python scripts/tts.py "Hello world"

---

Important notes

• Small model (0.6B) is the default. It uses less RAM and is faster. Switch to large (1.7B) for higher quality.
• CPU inference is slow. Expect 30-120 s per sentence for the 1.7B model. The 0.6B model is roughly 2x faster.
• Only one model variant is loaded at a time. Switching modes (e.g. custom-voice to voice-clone) triggers a model swap.
• Output .wav files land in tts_output/ by default.
• Models are downloaded to {baseDir}/models/ by default. Run download_models.py --size all to pre-download both sizes for offline use.
• Voice Design mode has no 0.6B variant — it always uses the 1.7B model regardless of QWEN_TTS_MODEL_SIZE.