Whisper ASR

update this Whisper ASR is a personal assistant that converts voice memos into text. The record and transcribe features of Whisper ASR let you get accurate and understandable transcriptions with just a few button presses. Currently, only 30-second audio chunks are supported, but that will change in the future. for better reflect latest work Whisper ASR is an offline speech recognition app that converts voice into text entirely on-device — no internet required. Record up to 30 seconds of audio and get accurate transcriptions with just a tap. Choose from three Whisper models: • tiny.en — bundled in the app, works instantly offline • small.en — higher accuracy English model, downloaded on first use • small — multilingual model supporting multiple languages Features: • Fully offline inference powered by quantized TFLite models • Audio playback with seek bar to review your recordings • Real-time recording timer • Copy transcription to clipboard • Models cached locally after first download Longer audio support coming in a future update.