Supertonic 3 voice previews
Copyright (c) 2026 Supertone Inc.

These recordings are machine-generated audio from the Supertonic 3 model.
Model revision: 724fb5abbf5502583fb520898d45929e62f02c0b
Model source: https://huggingface.co/Supertone/supertonic-3/tree/724fb5abbf5502583fb520898d45929e62f02c0b
Model license: BigScience Open RAIL-M, dated August 18, 2022.
The complete license, including Attachment A use restrictions, is provided in
LICENSE.txt at /agent-voices/supertonic/LICENSE.txt. Model use is subject to
that license and its use restrictions. This attribution does not imply
endorsement by Supertone Inc.

The twenty English and Russian previews were generated with the Harvest
Supertonic adapter at commit e9be4706f7f2bfbbff59eda0b3a47cb5a16668eb.
Provenance: har172-voice-proof/catalog-20261012/manifest.json
Source manifest SHA-256:
e9413af878a790b405050edfb1e7619e3e3b12e1337568ebb039c6179cde6c5b
The original 44,100 Hz mono 16-bit PCM WAV bytes are unchanged.
Per-recording text and SHA-256 hashes are recorded in the application's
libs/supertonic-previews.json catalog.

These are prerecorded local synthesis previews. They do not establish
availability on a meeting body or received audio acceptance in Google Meet.
No model weights are included in these preview assets.