Skip to content

DirectoryTools

MisoTTS logo

MisoTTS

Miso Labs' open-weights 8B text-to-speech model for building expressive English voice agents.

Visit MisoTTS

Audio & VoiceFreeWe checked the prices on misolabs.ai on 7 Oct 2026.

MisoTTS homepage screenshot

What is MisoTTS?

Miso Labs' open-weights 8B text-to-speech model for building expressive English voice agents.

Audio & VoiceModels & Open WeightsModelOpen weightsText to speech

Who is MisoTTS best for?

MisoTTS is best for developers building English voice agents who want expressive speech from a model they can run themselves.

  • Voice agents that should sound human

    It reads the user's audio as well as the text, so a reply can match the tone of the person speaking.

  • Fast spoken replies

    Miso Labs puts its response time at 110ms, against 700ms for ElevenLabs on its own chart.

  • Cloning a voice

    A ten-second audio clip is enough for the model to keep speaking in that voice.

  • Keeping voice data in-house

    The open weights are built for local deployment, so sensitive voice data can stay on your own servers.

How much does MisoTTS cost?

MisoTTS is free to download and run under a modified MIT license; Miso Labs lists its hosted API as coming soon and offers on-premises hosting and support contracts to enterprise teams on request.

Open weights$0Modified MIT license
  • 8B model on Hugging Face and GitHub
  • English only
  • You run it on your own hardware

Miso Labs' hosted API is listed as coming soon, and on-premises hosting with support contracts is available to enterprise teams on request. The license is MIT with one change: products with more than 50 million monthly users or more than $10 million in monthly revenue must show "Miso Labs" in their interface. Full pricing on misolabs.ai. We checked these prices on .

How do readers rate MisoTTS?

No one has rated MisoTTS yet. Be the first.

No ratings yetThe score and its stars show once 3 readers have rated MisoTTS.
Your rating

Sign in to rate. One rating per account.

Your rating of MisoTTS

What are the best MisoTTS alternatives?

Three tools on Recaply do MisoTTS's job. Chatterbox Turbo and ZONOS2 are open source, so you can run them on your own server.

Questions about MisoTTS

What is MisoTTS?

Miso Labs' open-weights 8B text-to-speech model for building expressive English voice agents.

Who is MisoTTS for?

MisoTTS is best for developers building English voice agents who want expressive speech from a model they can run themselves.

Is MisoTTS free?

Yes. MisoTTS has a free Open weights plan: 8B model on Hugging Face and GitHub.

Is there an open source alternative to MisoTTS?

Yes. Chatterbox Turbo and ZONOS2 are open source tools that do the same job, so you can run your own.

What are the best MisoTTS alternatives?

The closest alternatives on Recaply are Chatterbox Turbo, ZONOS2 and Fish Audio S2 Pro.

Where can I try MisoTTS?

On its official website, misolabs.ai.

Is Miso One the same as MisoTTS?

Yes. Miso Labs' co-founder introduced the 8B voice model as Miso One, while the company's website, GitHub and Hugging Face pages call the same model MisoTTS, or Miso TTS 8B.

What languages does MisoTTS support?

English only for now. The model also handles single conversation turns rather than full back-and-forth, and it cannot speak while the other person is talking, which Miso Labs lists as future work.