BaoBaoGo HSK 6 original learning text, locally synthesized.
Model: icefall AISHELL-3 VITS, speaker SSB0005/sid 0 (Apache-2.0).
https://huggingface.co/csukuangfj/icefall-tts-aishell3-vits-low-2024-04-06
Runtime: k2-fsa/sherpa-onnx (Apache-2.0). https://github.com/k2-fsa/sherpa-onnx
Dataset: AISHELL-3, Beijing Shell Shell Technology Co., Ltd. (Apache-2.0). https://www.openslr.org/93/
Explicit per-sense phonemes; word recordings with the same character can have distinct readings.
Synthetic speech, not copied from HearMandarin. No claim of expert validation of every clip.

MP3 encoding: lameenc 1.9.0 (LGPL-3.0-or-later wrapper) / LAME; local authoring tool only, encoder not distributed in app. https://pypi.org/project/lameenc/1.9.0/