Whisper Web logo

Whisper Web

Private speech-to-text and MP4-to-MP3 processing in your browser

Whisper Web Introduction

Whisper Web turns audio, video and microphone recordings into editable text on the user's device. Choose Tiny, Base or Small Whisper models, use WebAssembly or WebGPU where supported, and export TXT, JSON, SRT or VTT. The standard workflow supports media up to 300 MB and 20 minutes. A separate local workflow handles files up to 1 GB or one hour. It also converts one MP4, MOV or WebM file to MP3 locally. The site still needs network access for application assets and model downloads, and it is an independent tool rather than an official OpenAI service.

Alternative tools