← Back to feed

Speech-to-text in the browser with Whisper and Parakeet, no server

AI · Technology · 2026-10-26 10:53

A browser-based audio and video transcription tool built on WebGPU and WebCodecs, featuring OPFS caching, manual model downloads, real-time microphone recognition with VAD, and ONNX-quantized Whisper and Parakeet models—all running without a server.

Source: whisper.tanpreet.xyz

Auto-generated summary for reference only. RSS