Whisper Web

What is Whisper Web

Whisper Web is a browser-based transcription app for turning audio and video files into text without sending the media to a server. It is built around local processing, so the selected file is handled on your device while the page itself remains a web app you open in a browser.

The app supports transcription in 99+ languages and can produce editable text, timestamps, and subtitle files. It also includes a related MP4 to MP3 workflow for extracting audio from video when you only need the sound track.

The core use case is simple: open the site, provide media, choose a model and language setting, then export the result. The app is positioned for private transcription workflows where avoiding media uploads matters.

Key features

• Local speech recognition in the browser, with processing performed on the device rather than on Whisper Web servers.
• Supports audio and video inputs, including MP3, MP4, M4A, WAV, OGG, OPUS, WebM, AAC, and FLAC.
• Handles files up to 300 MB and 20 minutes in the standard transcription flow.
• Offers a larger-file path for media up to 1 GB and 1 hour.
• Exports to TXT, SRT, VTT, and JSON.
• Lets you choose from Whisper Tiny, Whisper Base, and Whisper Small.
• Includes audio language selection across a very large list of languages, plus output options for the same language or English translation.
• Provides two local compute modes: WebGPU and WebAssembly, with Automatic set to try WebGPU first.
• Accepts media by file picker, direct link, or microphone input.
• Includes a separate MP4 to MP3 utility for extracting audio without uploading the video.

How to use it

1. Open Whisper Web in a browser that supports the required local compute mode.
2. Choose your input method: drop a file, pick a file, paste a direct media URL, or record from the microphone.
3. Select the transcription model that fits your device and accuracy needs.
   • Tiny is the smallest download.
   • Base is the balance option.
   • Small adds more detail but uses more device resources.
4. Set the audio language if you know it, or review the available language list before starting.
5. Pick an output language if you want the transcript kept in the original language or translated to English.
6. Start transcription and wait while the browser processes the media locally.
7. Review the text, then export the result as TXT, JSON, SRT, or VTT.

If your file is longer than 20 minutes or larger than 300 MB, use the large-file transcription workflow instead of the standard one.

Who it is for

Whisper Web fits people who want a private transcription workflow without account creation or upload-based processing. It is especially relevant for:

• Journalists and researchers handling interviews or recorded conversations.
• Teams transcribing meetings from saved recordings without adding a bot participant.
• People converting voice memos into notes or timed text.
• Podcasters creating text, captions, or clip transcripts for short episodes and trailers.
• Anyone who needs local speech-to-text in a browser and wants control over exports.

It is also useful when you want subtitles or structured transcript output rather than just a plain text dump. Because the app can work with direct links and microphone input, it can fit both file-based and quick capture workflows.

What to know before you use it

Whisper Web is privacy-focused, but it is not fully offline in every sense. The app notes that the page, the Whisper model, direct media URLs, and analytics can still make network requests, even though the selected media and finished transcript stay in the browser.

There are also practical limits:

• The standard workflow is capped at 300 MB and 20 minutes.
• The large-file path goes higher, up to 1 GB and 1 hour, but it is a separate workflow.
• Direct links must open without sign-in and must allow the browser to read the media file.
• WebGPU may be faster, but it is only available on supported devices and browsers.
• WebAssembly is the compatibility fallback, so speed can vary by hardware.

Whisper Web does not replace an official recording archive or a legal transcript source when exact wording, compliance, or chain-of-custody matters. For critical records, verify the transcript against the original audio or video, and use the source file or an official transcription service when required.

For best results, choose the right model for your device, confirm the language before starting, and use the large-file transcription guide when your recording exceeds the standard limits. Whisper Web is strongest as a local browser transcription tool, especially when you want privacy, subtitle exports, and no media upload.