Unified result type for ASR transcription across all providers Named ASRTranscriptionResult to avoid conflict with MeetingTranscriptionService.Transcr

TranscriptionProvider implementation using transcribe.cpp for Whisper GGUF models.

Minimal driver that replays a `flagsChanged` / `keyDown` sequence through the pure `ModifierOnlyShortcutFlagsDecision` state machine. `nextPressed` is

Serializes final audio-engine releases and provides a completion barrier. `-[AVAudioEngine dealloc]` may wait for AVAudioIOUnit's internal queue. The

Result of a transcription operation

Owns the last strong reference to an audio engine while it is waiting to be released on the dedicated retirement queue. The token is intentionally se

Snapshot of the modifier-only tracking state fed into `ModifierOnlyShortcutFlagsDecision`.

One speaker-attributed portion of a file transcription.

A comprehensive speech recognition service that handles real-time audio transcription. This service manages the entire ASR (Automatic Speech Recognit

Central theme definition for the Fluid app. All colors, spacings and materials should be defined here to keep styling consistent and easy to evolve.

A TranscriptionProvider that uses Apple's new SpeechAnalyzer API (macOS 26+). This provides advanced speech-to-text with streaming capabilities.

A TranscriptionProvider that uses Apple's native SFSpeechRecognizer. This uses Apple's system speech path and lets macOS choose local or online recogn

TranscriptionProvider implementation using FluidAudio (optimized for Apple Silicon) This wraps the existing FluidAudio-based ASR for use on Apple Sili

A robust downloader for Hugging Face models with progress tracking and error handling. Supports downloading entire model repositories with proper file

Bridges provider callbacks into dependencies whose progress handlers are `@Sendable`. The callback is immutable; callers remain responsible for hoppin