Intended use and voice licenses
Commercial useVerifiedPiper's voice docs say Piper is intended for personal use and TTS research, that it adds no restrictions of its own on voice models, and that some voices have restrictive licenses. Check each voice's MODEL_CARD.
The `MODEL_CARD` file for each voice contains important licensing information. Piper is intended for personal use and text to speech research only; we do not impose any additional restrictions on voice models. Some voices may have restrictive licenses, however, so please review them carefully!
Fetched 2026-10-22 · last check 2026-10-22 · Source
Ownership of outputs
Ownership of outputsUnverifiedUnverified. No verbatim quote is stored for this row.
Fetched 2026-10-22 · last check 2026-10-22 · Source
Engine license: GPL-3.0-or-later
License caveatVerifiedThe piper-tts package is GPL-3.0-or-later. That governs the engine's code if you ship or modify it, not the audio you generate.
License GPL-3.0-or-later Requires Python >=3.9
Fetched 2026-10-22 · last check 2026-10-22 · Source
Example: ryan voice is non-commercial
License caveatVerifiedThe US English "ryan" voice's model card lists its dataset as CC BY-NC-SA 4.0 (non-commercial).
* URL: https://www.kaggle.com/datasets/roholazandie/ryanspeech * License: CC BY-NC-SA 4.0
Fetched 2026-10-22 · last check 2026-10-22 · Source
Example: lessac voice uses a research licence
License caveatVerifiedThe US English "lessac" voice's model card points its dataset license at the Blizzard 2013 Lessac research licence. Several other voices are fine-tuned from lessac.
* License: https://www.cstr.ed.ac.uk/projects/blizzard/2013/lessac_blizzard2013/license.html ## Training Trained from scratch.
Fetched 2026-10-22 · last check 2026-10-22 · Source
Lessac dataset licence excludes commercial use
License caveatVerifiedThe Blizzard 2013 Lessac licence excludes using its Materials (recordings, text and data) for any commercial purpose. It doesn't say whether that reaches a voice trained on the data. We treat lessac and its fine-tunes as not cleared to ship.
excludes loading, executing, storing, transmitting, displaying, copying, reverse engineering, developing, adapting, amending or otherwise using the Materials for any commercial purpose
Fetched 2026-10-22 · last check 2026-10-22 · Source
Example: libritts voice is CC BY 4.0
License caveatVerifiedThe US English "libritts" voice (904 speakers) lists its dataset as CC BY 4.0, which allows commercial use with attribution.
* URL: http://www.openslr.org/60/ * License: CC BY 4.0
Fetched 2026-10-22 · last check 2026-10-22 · Source
Example: joe voice data is CC0, fine-tuned from lessac
License caveatVerifiedThe US English "joe" voice lists CC0 data, but it was fine-tuned from the lessac voice, whose dataset licence is research-only.
* License: CC0 ## Training Finetuned from U.S. English lessac voice (medium quality).
Fetched 2026-10-22 · last check 2026-10-22 · Source
Steam AI disclosure
Steam disclosureVerifiedPlayer-facing art, sound, narrative, or localization made with this tool is the kind of shipped content Steam's Content Survey asks about (Pre-Generated). This is Valve's rule, not a statement from the vendor.
Pre-Generated: Any kind of content that ships with your game and is consumed by players that is created with the help of AI tools during development.
Fetched 2026-10-22 · last check 2026-10-22 · Source