Audio / Video to Text
Turn audio files, video files, microphone speech or browser tab audio into text with Whisper in your browser. Click timestamps to seek; export SRT or VTT.
Still loading. If the tool doesn’t appear, reload this page.
About this tool
Audio / Video to Text
Transcribe audio files, video files, live microphone speech, or browser tab audio into timestamped text with Whisper running in your browser, plus waveform sync.
Audio / Video to Text turns spoken audio and video tracks into clean, timestamped transcripts directly inside your browser. Upload MP3, WAV, M4A, OGG, FLAC, MP4, MOV, or WEBM files—or stream live microphone and screen/tab video audio—and watch words appear as they are recognised via Whisper (WebGPU with automatic WebAssembly fallback) and live browser speech recognition. Every audio track is decoded locally to 16 kHz PCM with an interactive waveform timeline so clicking any transcript segment seeks your media player to that moment, and you can export ready-to-use TXT, SRT, VTT, Markdown, or JSON files. Files and Whisper transcription stay in your browser (the Whisper model and helper libraries download from public CDNs the first time, and the FFmpeg decoder from unpkg only for formats the browser cannot read); the Browser Live Speech option uses your browser's own speech service, which in Chrome and Edge sends the microphone audio to the browser maker's servers. It downloads the Transformers.js library from cdn.jsdelivr.net for Whisper transcription; your media stays in your browser. It downloads the Whisper speech model from huggingface.co on first use; the transcription runs in your browser. It downloads the FFmpeg engine from unpkg.com when a file needs converting before transcription; your media stays in your browser.
How to use Audio / Video to Text
- 1
Choose an input mode—Upload Audio/Video for MP3, WAV, M4A, MP4, MOV, or WEBM files, Live Mic for speech dictation, or Screen / Tab Audio to transcribe playing video.
- 2
Select your transcription engine, language (auto-detect or 14+ languages), and task mode (Transcribe or Translate to English) in the Transcription Settings bar.
- 3
Start live capture or click Transcribe Audio/Video to watch timestamped text stream word by word alongside the interactive waveform and media player.
- 4
Click any timestamp badge to jump to that moment in the audio or video, edit segments inline, search keywords, and copy or export as TXT, SRT, VTT, Markdown, or JSON.
Why use Audio / Video to Text
Streaming output
Both Live Mic streaming and Whisper AI file decoding stream incremental words and timestamped segments as they are generated, rather than waiting silently until the end.
Audio, video, mic & tab capture in one workspace
Drop local audio/video files, dictate with a live voice level meter, or capture audio from another browser tab or screen video; Whisper keeps that audio in your browser.
Interactive waveform & export-ready subtitles
Scrub the visual waveform, click any segment timestamp to seek playback, edit text inline, and download standard .srt, .vtt, .txt, .md, or .json files.
Common use cases
- Transcribing recorded interviews, podcasts, lectures, and Zoom or Google Meet video recordings into searchable notes and SRT/VTT captions.
- Dictating from your microphone or transcribing a playing browser tab video live, with timestamps on every segment.
- Translating non-English audio or video clips (such as Hindi, Spanish, French, or Japanese) directly into English transcripts in one pass.
Frequently asked questions
How does Audio / Video to Text transcribe files without uploading them?
For local files, your browser extracts and resamples the audio track to 16 kHz mono PCM using the Web Audio API (with an automatic FFmpeg WebAssembly fallback for formats like MKV or AVI) and runs OpenAI Whisper ONNX models inside a background Web Worker.
Why does the Whisper AI model take a moment to load the first time?
On your first run with a Whisper model (Tiny ~40 MB or Base ~75 MB), your browser downloads the quantized ONNX weights and caches them in local Cache Storage. Later sessions load the model from your browser cache instead of downloading it again.
Which audio and video formats are supported?
You can upload MP3, WAV, M4A, AAC, OGG, FLAC, WEBM, MP4, MOV, MKV, and AVI files, as well as record live microphone audio or capture live audio from a browser tab or screen share.
Can it transcribe Hindi or other non-English languages and translate to English?
Yes. Choose Whisper Tiny Multilingual or Whisper Base Multilingual, keep Language on Auto-detect or pick your language, and optionally switch Task to Translate to English to output English text directly.
What is the difference between Browser Live Speech and Whisper AI in Live Mic mode?
Live Mic starts on Whisper, which works across browsers (including Firefox) and runs on your device once the model is cached. Browser Live Speech is an option you choose: it uses your browser's SpeechRecognition API for interim words as you speak in Chrome, Edge, and Safari. In Chrome and Edge that browser speech service sends your microphone audio to Google's or Microsoft's servers, and the page says so next to the switch.