Flock Speech (flockspeech)
Copyright 2026 Flock Speech

Licensed under the Apache License, Version 2.0 (see LICENSE). This covers the package code and the Junco
acoustic model weights (src/flockspeech/models/encoder.onnx, decoder-int8.onnx).

Third-party components:

- Vocos vocoder (src/flockspeech/models/vocoder.onnx): converted from charactr/vocos-mel-24khz,
  Copyright (c) 2023 Charactr Inc., MIT License (see src/flockspeech/models/LICENSE-vocos).
- Training data: LJ Speech Dataset (Keith Ito and Linda Johnson), public domain.

Dependencies installed by pip, not included in this package:

- phonemizer (GPL-3.0) and espeak-ng (GPL-3.0, shipped by espeakng-loader) turn text into phonemes.
- onnxruntime (MIT) and numpy (BSD-3-Clause).
