Carey voice (CareNotes Morning Huddle briefing)
Runtime assets for CareNotes' on-device Morning Huddle voice briefing. The apps download these files on first use; nothing here contains patient data.
| File | Used by | What it is |
|---|---|---|
carey_voice.bin |
iOS (FluidAudio PocketTtsVoiceCloner.loadVoice) |
Pocket TTS voice embedding, 83 x 1024 float32 |
carey_reference.wav |
Android (sherpa-onnx GenerationConfig.referenceAudio) |
24 kHz mono reference clip the embedding was cloned from |
android/pocket-tts-int8/* |
Android (sherpa-onnx OfflineTtsPocketModelConfig) |
Mirror of sherpa-onnx-pocket-tts-int8-2026-01-26 |
Provenance
carey_reference.wavwas synthesized with Kokoro-82M (Apache 2.0), voiceaf_heart, byScripts/tts/make_carey_reference.pyin the CareNotes repo. It is not a recording of a person.carey_voice.binwas produced from that clip by FluidAudio 0.17.1 (fluidaudiocli tts --clone-voice ... --save-voice).- Pocket TTS is by Kyutai (https://huggingface.co/kyutai/pocket-tts), CC-BY 4.0. The Android
files are the k2-fsa sherpa-onnx ONNX export (https://github.com/k2-fsa/sherpa-onnx/releases/tag/tts-models);
see
android/pocket-tts-int8/LICENSE.
- Downloads last month
- -