Voxint
Copyright 2026 Voxint contributors

This product includes software developed as part of the Voxint project
(Apache License 2.0).

Voxint orchestrates third-party models and libraries that are licensed
separately; model weights are not stored in this repository but ARE
redistributed (sha256-pinned) via the project's model-asset releases and
container images:

- faster-whisper / OpenAI Whisper weights (MIT) — downloaded at build/run time
- pyannote speaker-diarization-3.1 pipeline, vendored via the
  pyannote-models-v1 asset release (services/pyannote/models/provenance.json):
  - pyannote/segmentation-3.0 (MIT) — (c) pyannote contributors
  - WeSpeaker ResNet34 speaker-embedding weights (CC-BY-4.0) — trained by the
    WeSpeaker project (github.com/wenet-e2e/wespeaker) on VoxCeleb, repackaged
    for pyannote.audio by Herve Bredin; redistributed with attribution
- NVIDIA NeMo TitaNet-Large (CC-BY-4.0 model license) — vendored via the
  titanet-onnx-v1 asset release with attribution

Consult each upstream project for its exact license terms.
