Grow Word speech recordings and pronunciation repairs Most saved speech uses the project's existing Kokoro bf_emma recordings. No speech service is contacted during play. The following three complete human recordings are by speaker and recorder Back ache, from Lingua Libre, distributed through Wikimedia Commons: am Runtime: assets/reviewed-audio/word-am.mp3 Source: https://commons.wikimedia.org/wiki/File:LL-Q1860_(eng)-Back_ache-am.wav an (stressed form) Runtime: assets/reviewed-audio/word-an.mp3 Source: https://commons.wikimedia.org/wiki/File:LL-Q1860_(eng)-Back_ache-an.wav year Runtime: assets/reviewed-audio/word-year.mp3 Source: https://commons.wikimedia.org/wiki/File:LL-Q1860_(eng)-Back_ache-year.wav Speaker profile: https://lingualibre.org/wiki/Q113883 Recorder: https://commons.wikimedia.org/wiki/User:Back_ache These three recordings and our adapted MP3 versions are licensed under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0): https://creativecommons.org/licenses/by-sa/4.0/ Changes: retained each complete recording, resampled to mono 24 kHz, lowered its volume to match the surrounding voice, applied 5 ms tapers at the quiet edges, added 20/40 ms of silence, and encoded as MP3. No words or speech sounds were removed or added. Credit does not imply endorsement. This licence covers these three audio assets; the game code and other assets retain their separate licences. Sixteen spoken meaning examples added 30 October 2026: original TylerBuilds sentence/question prompts, saved Kokoro bf_emma recordings generated locally as complete utterances. No timestamp cropping. These invite shared conversation and are not an assessment. Model: hexgrad/Kokoro-82M, Apache-2.0.