Adds a persisted transcription language for OpenAI Whisper and local multilingual Whisper, plus an independent OpenAI transcription model setting.
Android also gains a dedicated offline French option based on Whisper Small fine-tuned on French Common Voice. Its ~375 MB int8 model downloads through the existing manager with pinned hashes, resumable downloads and independent deletion. Selecting it requests French without overwriting the shared language preference. Existing providers and defaults are preserved.
Model provenance, Apache license, reproducible export and an actual French audio smoke test are documented in docs/models/whisper-small-fr. Versioned assets are published in the public Gitea package registry.
Validation: JavaScript and Kotlin tests pass, including persistence, model selection, switching back, interrupted downloads and deletion isolation. Actual CPU inference passed using the same sherpa-onnx version as Android. On-device latency, RAM and glasses microphone checks remain pending. No firmware changes.
Adds a persisted transcription language for OpenAI Whisper and local multilingual Whisper, plus an independent OpenAI transcription model setting.
Android also gains a dedicated offline French option based on Whisper Small fine-tuned on French Common Voice. Its ~375 MB int8 model downloads through the existing manager with pinned hashes, resumable downloads and independent deletion. Selecting it requests French without overwriting the shared language preference. Existing providers and defaults are preserved.
Model provenance, Apache license, reproducible export and an actual French audio smoke test are documented in docs/models/whisper-small-fr. Versioned assets are published in the public Gitea package registry.
Validation: JavaScript and Kotlin tests pass, including persistence, model selection, switching back, interrupted downloads and deletion isolation. Actual CPU inference passed using the same sherpa-onnx version as Android. On-device latency, RAM and glasses microphone checks remain pending. No firmware changes.
Blendman974
changed title from feat: language selection for OpenAI and on-device Whisper to Configure Whisper languages and OpenAI transcription model2026-11-07 23:37:24 +00:00
Blendman974
changed title from Configure Whisper languages and OpenAI transcription model to Voice: transcription languages, model selection and offline French Whisper2026-11-10 22:23:54 +00:00
Some required checks failed
Secret scan / gitleaks (pull_request) Canceled after 0s
Tests / test (pull_request) Canceled after 0s
Tests / stats-server (pull_request) Canceled after 0s
Tests / evenhub-extensions (pull_request) Canceled after 0s
Tests / test-kotlin (pull_request) Canceled after 0s
Checking for merge conflicts…
View command line instructions
Checkout
From your project repository, check out a new branch and test the changes.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Adds a persisted transcription language for OpenAI Whisper and local multilingual Whisper, plus an independent OpenAI transcription model setting.
Android also gains a dedicated offline French option based on Whisper Small fine-tuned on French Common Voice. Its ~375 MB int8 model downloads through the existing manager with pinned hashes, resumable downloads and independent deletion. Selecting it requests French without overwriting the shared language preference. Existing providers and defaults are preserved.
Model provenance, Apache license, reproducible export and an actual French audio smoke test are documented in docs/models/whisper-small-fr. Versioned assets are published in the public Gitea package registry.
Validation: JavaScript and Kotlin tests pass, including persistence, model selection, switching back, interrupted downloads and deletion isolation. Actual CPU inference passed using the same sherpa-onnx version as Android. On-device latency, RAM and glasses microphone checks remain pending. No firmware changes.
feat: language selection for OpenAI and on-device Whisperto Configure Whisper languages and OpenAI transcription modelConfigure Whisper languages and OpenAI transcription modelto Voice: transcription languages, model selection and offline French WhisperView command line instructions
Checkout
From your project repository, check out a new branch and test the changes.