Speaks English subtitles in one of 34 named voices, on your own computer. Parler-TTS has no voice files: it follows a written description of the voice, and you can write your own. It doesn't clone voices.

What it does

34 named voices

Each voice is a speaker the model was trained on. The add-on asks for that speaker at a moderate pace, close to the microphone, with no background noise, so the same name gives the same voice on every line.

Or describe the voice

Write a description in the add-on's settings, such as a low-pitched female voice in a quiet room, and every Parler speaker follows it instead.

Slow on the CPU

Expect 10 to 30 seconds per line. The first line also waits for the model to download.

No cloning

It can't copy a speaker's voice from the video. For that, use Qwen3-TTS, Coqui XTTS or F5-TTS.

Voices

In the Speakers tab, choose Parler-TTS as a speaker's Dubbing engine, then pick a Voice. Six names are marked recommended: Jon, Lea, Gary, Jenna, Mike and Laura. Give each speaker a different name so they don't sound alike.

Each line comes out as long as the model makes it. Turn on Fit dub to subtitle duration for the speaker to stretch or squeeze it to its subtitle.

34 voices in 1 language.

en 34

Models

One model, downloaded the first time you generate a line. It goes into the standard Hugging Face cache (~/.cache/huggingface/hub, unless you set HF_HOME), outside the add-on's folder, so updating or removing the add-on keeps it. Delete it from there yourself to free the space.

Model	Download
parler-tts-mini-v1	2.2 GB

Options

The add-on reads these when it starts, so a change applies once it has been idle for five minutes, or after you restart Subtitld.

In Settings › Add-ons › Parler-TTS Mini v1 (offline, prompt-driven) › Configure.

Inference device

Default: cpu

Leave this on CPU. The released builds are made with PyTorch's CPU package, so CUDA won't work with them; MPS is marked beta.

Default voice description

Default: empty

Empty, each voice uses its own description. Text here replaces it for every Parler speaker in the project, whichever voice they picked. The built-in descriptions start with the speaker's name, which keeps a voice the same from line to line, so start yours the same way.

Languages

English only. The add-on doesn't read the project's language, so it reads every line as English.

en

Platforms

System	Download	Tested
Linux x86-64	436.3 MB	Built and started in CI
macOS (universal)	302.9 MB	Built and started in CI
Windows x86-64	284.3 MB	Built and started in CI

Each release is built on Linux, macOS and Windows, and CI checks that the program starts and lists its voices. CI doesn't generate speech.

Troubleshooting

internalThe model failed to load or to speak the line; the message says which. A failed download on first use ends here too: check your connection and free disk space, and try again. If you chose CUDA or MPS, set the device back to CPU.
bad_paramsThe subtitle has no text to speak. Give it text, or leave it out of the dub.

License

Add-on

Apache-2.0

Engine

Apache-2.0

parler-tts

Models

Apache-2.0

Parler-TTS Mini v1 weights

Fine for commercial work: the add-on, the parler-tts library and the model weights are all Apache-2.0.

Parler-TTS (offline, prompt-driven)

What it does

Voices

Models

Options

Languages

Platforms

Troubleshooting

License

Versions

Other Voices add-ons