Pocket TTS (offline, 7 languages)

What it does

Before you install

Pocket TTS asks little of your computer:

• Disk: about 220 MB for each language you dub in, downloaded on its first line.
• Memory: about 660 MB while it speaks, or 450 MB with Use less memory.
• Processor: on Intel and AMD, one with AVX2 and FMA, which Intel processors have had since 2013 and AMD since 2015. Some Celeron, Pentium and Atom processors lack them, and there the first line fails with a message that says so. Every Apple Silicon Mac works.

It does not clone voices. Pocket TTS can, but Kyutai gives its cloning model only to people who sign in to Hugging Face and accept its terms, so the add-on uses the model without it. To dub in the speaker's own voice, use VoxCPM2 or Qwen3-TTS.

Voices

In the Speakers tab, choose Pocket TTS as a speaker's Dubbing engine, then pick a Voice. Turn on Fit dub to subtitle duration for the speaker to stretch or squeeze each line to its subtitle. The project's language picks the model, so set it before you dub.

23 voices.

A voice may keep something of its accent in another language, so try a few in your project's language before you choose. Kyutai's own choices are Alba for English, Estelle for French, Giovanni for Italian, Lola for Spanish and Daan for Dutch. Each voice downloads the first time you use it, under 8 MB.

Kyutai offers four more voices, which the add-on leaves out: two come from recordings licensed for non-commercial use only, and two don't say where they come from.

Models

Pocket TTS has one model per language. None is inside the add-on: each downloads into Subtitld's shared model cache the first time you dub in its language, from Hugging Face at a pinned revision, each file checked. An interrupted download resumes where it stopped. Updating or removing the add-on keeps the models.

Model in the add-on's settings picks the size. Small is Kyutai's newest English model, with 18 million parameters instead of 100 million. Large preview loads Kyutai's 24-layer models, 670 MB per language and 1.3 GB for English, at about real time. Kyutai calls them previews, not final models. In our test they made no fewer mistakes than Standard.

Options

In Settings › Add-ons › Pocket TTS (offline, 7 languages) › Configure.

Languages

English, French, German, Portuguese, Italian, Spanish and Dutch. A project in any other language gets a message that says so: pick another dubbing engine for it.

• en
• fr
• de
• pt
• it
• es
• nl

Platforms

Linux x86-64, macOS on Apple Silicon and Windows x86-64. Each release is built on all three, and on each one CI installs the add-on as Subtitld does, downloads the English models and dubs with them. On Linux it needs glibc 2.34 or newer, as Subtitld itself does.

Troubleshooting

• network_unavailableThe model couldn't be downloaded. Check your connection and try again: the download resumes.
• disk_fullNot enough disk space for the model. Each language needs about 220 MB.
• unsupported_languageThe project's language is not one of the seven, or not set. Set it in the project, or pick another dubbing engine.
• unsupported_voiceThe project names a voice this version doesn't have. Pick the voice again in the Speakers tab.
• bad_paramsThe line has no text.
• internalSomething failed inside the add-on; the message has the details. If it says the processor lacks AVX2 and FMA, this computer can't run Pocket TTS.

License

Fine for commercial work: the add-on is Apache-2.0, Pocket TTS is MIT, and its models are Kyutai's, under CC-BY-4.0. The voices come from recordings under CC0 or CC-BY-4.0. The dubs you make are yours. Kyutai forbids using Pocket TTS to imitate a real person's voice without their consent, or to deceive.