twin_stream 0.2.0

Simultaneous speech-to-text and audio file recording in a single API. Solves the Android microphone lock issue that prevents using recording and speech recognition at the same time.

0.2.0

• 🧠 True Android Simultaneous STT Support: Added voskModelPath to TwinStreamConfig. When provided, the package leverages vosk_flutter and a custom raw PCM stream-to-WAV writer to fully bypass the Android OS Microphone Lock, allowing true simultaneous recording and offline transcription on strict devices.
• 🐛 Bug Fix: Addressed Issue #1 by clearing non-fatal STT timeout errors (such as error_speech_timeout) immediately after they are emitted, preventing the duration timer from rebroadcasting the same error to the UI continuously on fallback.

0.1.0

• 🎉 Initial release of twin_stream
• Simultaneous speech-to-text and audio file recording in a single API
• Android: Automatic microphone lock fallback (simultaneous → sequential)
• iOS / macOS: True simultaneous recording and speech recognition
• TwinStreamController — main API with start(), stop(), cancel(), pause(), resume()
• TwinStreamConfig — configurable sample rate, encoder, locale, Android strategy
• TwinStreamState — real-time state stream with live transcript, sound level, recording duration
• TwinStreamResult — session results with audio file path, transcript, confidence, alternates
• Custom exception hierarchy: TwinStreamPermissionException, TwinStreamMicrophoneLockException, TwinStreamInitializationException, TwinStreamRecordingException
• Three Android recording strategies: simultaneousWithFallback, sequential, simultaneousStrict
• Support for multiple audio encoders: WAV, AAC, Opus, FLAC, PCM
• Comprehensive documentation and example app