speech

speech

On-device speech recognition, with a speech_to_text-backed implementation. Part of the modular-flutter-template core modules.

Two packages:

• speech_api — the abstract contract: SpeechRecognizerApi, the sealed SpeechAvailability and SpeechResult. Feature code and other _api packages depend on this, never on speech_impl.
• speech_impl — the concrete implementation over speech_to_text, and SpeechModule. Only the app's composition root depends on this.

Using it

Workspace packages: depend on them by name, with no source:

dependencies:
  speech_api:
  speech_impl:   # the app's composition root only

The host app still owns the platform setup speech_to_text needs: the microphone and speech-recognition usage descriptions in Info.plist, and the RECORD_AUDIO permission in AndroidManifest.xml.

Local development

Resolved with the rest of the workspace: run flutter pub get at the repo root. Tests run with the others through dart run melos test.

Speech recognition needs a physical device; the iOS Simulator cannot do it reliably, so the tests here cover only the platform-free parts.