flutter_native_speech_to_text 1.3.2

Speech-to-text for Flutter by zakaria5729. Uses Android RecognizerIntent, Apple Speech, Windows, Vosk on Linux and the Web Speech API.

1.3.2

• Added a live web demo link (hosted on GitHub Pages) to the README.
• The example app now has a Bangla/English language switch.

1.3.1

• Added an author section to the README and credited the author in the package description.

1.3.0

• Windows: real speech recognition using the Windows speech recognizer (Windows.Media.SpeechRecognition), replacing the stub.
• Linux: speech recognition using Vosk and PulseAudio, loaded at runtime. It needs libvosk, libpulse-simple and a VOSK_MODEL_PATH folder of models; without them isAvailable() is false and listen() throws NO_RECOGNIZER.
• The Windows and Linux implementations have not been tested on those platforms yet.

1.2.0

• Added iOS and macOS support using Apple's Speech framework (SFSpeechRecognizer), shared between both platforms. The previous iOS stub is replaced.
• Apple platforms do not support Bengali; listen() throws LANGUAGE_NOT_SUPPORTED for unsupported languages.
• Added FlutterNativeSpeechToText.isLanguageSupported(language) on every platform.
• Added Windows and Linux native stubs: they never crash, isAvailable() and isLanguageSupported() return false, and listen()/stop()/cancel() fail with PLATFORM_NOT_SUPPORTED.
• Includes the web fixes from 1.1.1 below.

1.1.1

• Web: listen() now completes as soon as the browser reports an error, instead of waiting for an end event that Safari may never send.
• Web: stop() falls back to the latest result if the browser does not end the session within three seconds.
• Web: Brave is detected, since it blocks the speech service. isAvailable() returns false and listen() throws NO_RECOGNIZER with an explanatory message.
• Web: new SERVICE_NOT_ALLOWED error code, separate from PERMISSION_DENIED, for browsers that block their speech service (for example Safari when the requested language is not installed for Dictation).

1.1.0

• Added web support using the browser's Web Speech API (SpeechRecognition / webkitSpeechRecognition), exposed through the existing listen, stop, cancel and isAvailable API.
• New error codes on web: PERMISSION_DENIED, NO_MICROPHONE, NETWORK, LANGUAGE_NOT_SUPPORTED.
• Minimum SDK raised to Dart 3.4 / Flutter 3.22.

1.0.0

• Initial release of the plugin.
• Speech recognition using Android's native ACTION_RECOGNIZE_SPEECH Intent.
• Configurable parameters for RecognizerIntent extras:
  • languageModel (free form or web search)
  • language and languagePreference
  • prompt, maxResults
  • partialResults, confidenceScores
  • Speech silence length limits (speechInputMinimumLengthMillis, speechInputCompleteSilenceLengthMillis, speechInputPossiblyCompleteSilenceLengthMillis)
• Availability check with FlutterNativeSpeechToText.isAvailable().
• Stop and cancel session controls.
• iOS support scaffold: the iOS implementation is a safe stub that never crashes.
  • isAvailable() returns false.
  • listen() / stop() / cancel() throw SpeechToTextException with code PLATFORM_NOT_SUPPORTED, so host apps can display and log the error.
• Complete example application demonstrating integration.