whisper_cpp_flutter_plus library
Offline whisper.cpp transcription, streaming, recording, and VAD support.
Classes
- ModelDownloadProgress
- Byte-level progress for a model download.
- RecordingChunk
- A chunk of mono floating-point PCM audio.
- TranscribeOptions
- Options controlling a transcription job.
- VadSegment
- A speech interval returned by voice activity detection.
- WhisperAudio
- Utilities for preparing audio for whisper.cpp.
- WhisperBatchInput
- One input in a sequential transcription batch.
- WhisperBatchItemResult
- Completion information for one successfully processed or captured item.
- WhisperBatchProgress
- Ordered progress emitted by a sequential transcription batch.
- WhisperBatchTask
- Owns one sequential batch and its active inference.
- WhisperConfig
- Native model-loading options.
- WhisperEngine
- A loaded whisper.cpp model context used to run transcription jobs.
- WhisperModelCatalog
- Curated models suitable for common mobile speech workflows.
- WhisperModelDescriptor
- Immutable metadata for a checksum-pinned downloadable model.
- WhisperModelManager
- Stores, discovers, downloads, and removes whisper model files.
- WhisperPcmBatchInput
- A batch input containing prepared mono 16 kHz floating-point PCM.
- WhisperRecorder
- Captures mono floating-point PCM from the device microphone.
- WhisperResult
- Completed transcription output.
- WhisperSegment
- A contiguous transcription segment.
- WhisperStreamConfig
- Controls how often live audio is decoded and how long text remains provisional before it can be confirmed.
- WhisperStreamTask
- Owns one continuous transcription and its audio subscription.
- WhisperStreamUpdate
- A display-ready snapshot of a live transcription.
- WhisperTask
- Handle for one asynchronous transcription job.
- WhisperToken
- A decoded whisper.cpp token and its timing/probability metadata.
- WhisperVad
- Runs whisper.cpp voice activity detection on mono 16 kHz PCM.
- WhisperWavBatchInput
- A batch input containing a WAV file decoded before inference.
Enums
- WhisperLogLevel
- Verbosity of native whisper.cpp diagnostic logging.
- WhisperModelLanguageScope
- Language coverage of a catalog model.
- WhisperModelPurpose
- Intended use of a catalog model.
- WhisperPerformanceMode
- Reusable offline transcription profiles with different performance goals.
- WhisperSamplingStrategy
- Decoding algorithm used to select transcription tokens.
Extensions
- WhisperResultExport on WhisperResult
- Text and subtitle exports for a completed transcription.
Typedefs
-
WhisperBatchInferenceStarter
= Future<
WhisperResult> Function(Float32List samples, TranscribeOptions options) - Starts one inference pass for a batch input.
-
WhisperBatchSampleLoader
= Future<
Float32List> Function(WhisperBatchInput input) - Loads the samples for one batch input.
Exceptions / Errors
- WhisperException
- An error reported by the Dart wrapper or native whisper.cpp runtime.