AdAdvertise with us·Reach an audience of professional tech enthusiasts
Favicon of Kokoro TTS website

Kokoro TTS

Kokoro is a lightweight open-weight text-to-speech model with 82 million parameters. Its Apache-licensed weights can be run locally via a Python package or deployed in production and commercial projects.

Price:
Free
Alternative to:
Favicon of Speechify website Speechify
Favicon of ElevenLabs website ElevenLabs
Screenshot of Kokoro TTS website

About Kokoro TTS

Kokoro TTS is an open-weight text-to-speech model with 82 million parameters, built on the StyleTTS 2 and ISTFTNet architectures. It is aimed at developers and hobbyists who want to generate speech from text on their own hardware or inside their own products, without paying per character to a hosted service.

Key benefits of Kokoro TTS include:

  • Apache 2.0 license: The weights can be used in personal projects, production environments and commercial APIs.
  • Small and fast: At 82M parameters it is designed to be faster and cheaper to run than larger TTS models.
  • Multiple languages and voices: Version 1.0 ships with 8 languages and 54 voices.
  • Simple Python usage: Install the kokoro package with espeak-ng and generate 24 kHz audio in a few lines of code, including on Google Colab.
  • Permissive training data: It was trained on public domain, permissively licensed and synthetic audio, with no custom voice clones.

If you are leaving ElevenLabs or Speechify, Kokoro lets you produce speech locally for free, and hosted APIs serving it are cited at under $1 per million characters. The trade-off is that it is a model, not a polished app: you need some comfort with Python or a third-party front end, and the model page does not mention voice cloning or the reading and editing features those services offer. Beware of unofficial websites with "kokoro" in their domain, which the author says are not affiliated.

Share:

Stay in the loop

Weekly roundup of new products. No spam, unsubscribe anytime.

Trusted by 735+ people

Similar alternatives