Skip to content

Add experimental multilingual speech support - #3

Draft
edsonlima2000 wants to merge 1 commit into
mainfrom
agent/multilingual-speech
Draft

Add experimental multilingual speech support#3
edsonlima2000 wants to merge 1 commit into
mainfrom
agent/multilingual-speech

Conversation

@edsonlima2000

Copy link
Copy Markdown
Owner

What changed

  • adds persistent speech-language selection for Portuguese, English, and Spanish
  • centralizes language-specific wake, ending, cancellation, and service commands
  • keeps Portuguese Vosk bundled and downloads official English/Spanish small models on demand
  • pins SHA-256 validation for optional Vosk model downloads
  • passes the selected language dynamically to the local Whisper.cpp JNI runtime
  • documents model, privacy, licensing, and future-language constraints

Why

Prepare SmartGlassesNotes for an international community without mixing languages in one recognizer or weakening the working Portuguese trigger.

User impact

Portuguese remains the default. English and Spanish are explicitly experimental and require a one-time model download when first selected.

Validation

  • assembleDebug
  • test
  • lint
  • verified official model archive roots and SHA-256 hashes

Still required

Physical wake/audio validation by fluent English and Spanish speakers. Per owner request, this APK was not installed on the connected phone.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant