TTS

What it is and what it does

TTS is a PyTorch-based library for neural text-to-speech synthesis that provides pretrained models and architectures including Tacotron2, Glow-TTS, VITS, XTTS, Bark, and Tortoise. It handles the full pipeline from text input to audio output, supporting both inference with released models and training custom models on new datasets.

The library depends on a substantial stack of 39 runtime packages: torch, torchaudio, scipy, librosa, scikit-learn, and language-specific tools including jieba, g2pkk, bangla, jamo, hangul-romanize, gruut, and pysbd. It includes vocoder models (MelGAN, HiFiGAN, ParallelWaveGAN) to convert spectrograms to waveforms, speaker encoders for multi-speaker synthesis, and utilities for dataset curation. Installation is straightforward via pip, though the dependency footprint and model downloads make it a medium-friction package.

Use it for:

• Generate speech from text in applications requiring voice output, using pretrained models without training.
• Train custom TTS models on proprietary voice datasets for domain-specific or branded voice synthesis.
• Implement voice cloning by fine-tuning existing models with speaker-specific audio samples.
• Build multilingual speech synthesis pipelines with language-specific models and tools.
• Integrate TTS into web services or applications via the library's Flask-based server for synthesis.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

TTS is a deep learning library for text-to-speech synthesis that generates spoken audio from text using pretrained models across multiple languages, with support for model training and fine-tuning.

Yes, with conditions. Install if you need neural TTS synthesis and can accept the MPL-2.0 copyleft constraint and substantial dependency footprint. The library has recent commits and is in active use, but has not released since 2023-12-12. Avoid if you require active maintenance guarantees or proprietary licensing flexibility.

Install

tts on PyPI

pip install tts

uv add tts

poetry add tts

Installing TTS

Before you install

Medium install friction due to 39 runtime dependencies including torch, torchaudio, scipy, and librosa. The project is dormant with last commit on 2024-08-16. Prebuilt wheels available for Python 3.9–3.11 on Linux x86_64 reduce friction for standard environments.

License in practice

Licensed under MPL-2.0, a copyleft license requiring derivative works to be distributed under the same license and disclosing source code modifications. This affects proprietary deployments.

Quickstart

pip install TTS

from TTS.api import TTS
tts = TTS(model_name="tts_models/en/ljspeech/glow-tts", gpu=False)
tts.tts_to_file(text="Hello world", file_path="output.wav")

Requires Python >=3.9.0, <3.12. Large model downloads on first use; GPU recommended for inference speed but CPU inference supported.

Verify before relying

• Whether dormant status indicates active maintenance or abandonment despite recent commits.
• Performance characteristics and inference latency for different model architectures.
• Memory requirements for different pretrained models during inference and training.
• Actual number of supported languages and models available in the pretrained collection.

Package facts

Evidence: TTS-0.22.0-cp310-cp310-manylinux1_x86_64.whl; TTS-0.22.0-cp311-cp311-manylinux1_x86_64.whl; TTS-0.22.0-cp39-cp39-manylinux1_x86_64.whl

Tags

More Software Development packages