Sign in Submit site

VoiceCraft

github.com/jasonppy/VoiceCraft AI Speech Synthesis Open-source project

VoiceCraft supports zero-shot speech editing and text-to-speech generation.

Visit VoiceCraft 24 alternatives
github.com/jasonppy/VoiceCraft
VoiceCraft homepage screenshot

About VoiceCraft

VoiceCraft is a GitHub repository for zero-shot speech editing and text-to-speech (TTS) generation. According to the project README, it is a token infilling neural codec language model. The README states that it performs well on speech editing and zero-shot TTS using in-the-wild data such as audiobooks, internet videos, and podcasts. To clone or edit a voice it has not been trained on, it needs only a few seconds of reference audio.

The project can be run in several ways: in Google Colab, with Docker, from a local installation, or as a standalone command-line script. The Gradio interface is also available on Hugging Face Spaces. Model weights are published on Hugging Face. The README notes a length limit for one fine-tuned TTS model: the combined prompt and generation length must be 16 seconds or less. Training and fine-tuning are described in the README, but the project lists better guidance on these as a to-do item.

Key features

  • Speech editing through token infilling with a neural codec language model
  • Zero-shot text-to-speech from a few seconds of reference audio
  • Pretrained TTS models in 330M and 830M parameter sizes
  • Gradio web interface runnable locally, in Colab, or on Hugging Face Spaces
  • Command-line scripts for TTS and speech editing with adjustable inputs and settings
  • Docker setup tested on Linux and Windows
Best for: Researchers and developers who want to experiment with speech editing or text-to-speech models and are comfortable setting up Python environments or using Colab.
Good for
  • Several run options, including Google Colab, Docker, local installation, and standalone scripts
  • Requires only a few seconds of reference audio to clone or edit an unseen voice
  • The README describes state-of-the-art results on in-the-wild audio, including audiobooks, internet videos, and podcasts
Watch out for
  • For the fine-tuned giga330M TTS model, the combined prompt and generation length must not exceed 16 seconds
  • Docker setup on Linux requires the NVIDIA container toolkit
  • The README lists better guidance on training and fine-tuning as a to-do item

VoiceCraft at a glance

Type
Open-source project
Pricing
Not confirmed. Check the site
Open source
Not stated
Language
English
Last checked
Oct 6, 2026 · Online when checked
Listed
Oct 5, 2026
speechspeech synthesis

VoiceCraft alternatives

Tools and sites you could use instead, ranked by how closely they match.

All 24 alternatives

Popular in AI Speech Synthesis

Well-known listings from the same topic.

All 51
Speechifyspeechify.comSPEECHMAspeechma.comElevenReaderelevenreader.ioVanillaVoicevanillavoice.comPaper2Audiopaper2audio.com
Related topicsAI Business ToolsAI Image GeneratorsAI Infrastructure & MLOpsAI Agents & AutomationAI Data & AnalyticsAI Chatbots & AssistantsAI Voice & AudioAI Writing Tools

VoiceCraft FAQ

How much reference audio is needed to clone a voice?
According to the README, VoiceCraft needs only a few seconds of reference audio to clone or edit an unseen voice.
Is there a length limit for generated speech?
For the fine-tuned giga330M TTS model, the README says the maximal prompt plus generation length should be 16 seconds or less.
Can I try VoiceCraft without installing anything?
The README describes Google Colab as the simplest way to try speech editing or TTS inference.
Is something wrong or out of date?

Tell us and an editor will check it.

Own this website?

Claim the listing and suggest updates yourself. An editor checks every claim first.

Claim this listing

Listing maintained by Chris McClure and the WebIndexly editors · Updated Oct 13, 2026 · How we rank