Get up and running with TTS Studio in 5 minutes.
git clone https://github.com/bryanstevensacosta/tts-studio.git
cd tts-studio/apps/core
./setup.sh
source venv/bin/activatefrom api.studio import TTSStudio
# Initialize
studio = TTSStudio()
# Create profile from audio samples
profile = studio.create_voice_profile(
name="my_voice",
sample_paths=[
"./data/samples/neutral_01.wav",
"./data/samples/happy_01.wav"
],
language="es"
)
print(f"Profile created: {profile['profile']['id']}")# Generate audio from text
result = studio.generate_audio(
profile_id="my_voice",
text="Hola, esta es una prueba de mi voz clonada.",
output_path="./data/outputs/output.wav"
)
print(f"Audio generated: {result['output_path']}")from api.studio import TTSStudio
from pathlib import Path
# Initialize
studio = TTSStudio()
# Validate samples first (optional)
validation = studio.validate_samples([
"./data/samples/neutral_01.wav",
"./data/samples/happy_01.wav"
])
if validation["all_valid"]:
# Create profile
profile = studio.create_voice_profile(
name="my_voice",
sample_paths=[
"./data/samples/neutral_01.wav",
"./data/samples/happy_01.wav"
],
language="es"
)
# Generate audio
result = studio.generate_audio(
profile_id=profile["profile"]["id"],
text="Hola, ¿cómo estás?",
output_path="./data/outputs/greeting.wav"
)
print(f"✓ Audio generated: {result['output_path']}")
else:
print("✗ Some samples are invalid")For best results:
- Duration: 3-30 seconds per sample
- Quantity: 1-3 samples
- Quality: Clear audio, minimal background noise
- Format: WAV, 12000 Hz, mono, 16-bit (or will be converted)
- Configuration Guide - Configure TTS Studio
- Recording Samples - Learn how to record quality samples
- API Reference - Explore the full API