Transcribe from a link, without downloading anything

The detour you can skip

The usual route to a transcript of something published online has four steps: find a downloader, download the video, convert it to audio, upload the audio. Three of those steps exist only to move bytes around, and each one is a chance to end up with a 900 MB file you did not want.

Paste the address instead. WhoScribe reads the page, pulls the audio track and queues it like any uploaded file.

What happens after you paste

The link is checked immediately — you find out within seconds whether the page has usable audio and roughly how long it is. That check is quick because it only reads metadata.

The actual fetch happens in the background, on the queue. Nothing keeps you waiting on the page; close the tab and the transcript will be in your workspace when it is ready.

From there it is an ordinary transcript: speaker labels, inline editing, translation, subtitle export. Nothing about it is second-class because it arrived by link.

Where it works, and where it does not

It works on pages that serve a normal audio or video stream — most public video hosts, podcast episode pages, direct MP3 and MP4 addresses.

It does not work on anything behind a login, a paywall or a members-only area. The importer visits the page as an anonymous visitor, so if you cannot see the media while logged out, neither can it. For those, download the file yourself and upload it.

Very long streams and live broadcasts are also poor candidates: a live URL has no fixed end, and there is nothing to fetch until it does.

One thing to keep in mind

Being able to fetch a recording is not the same as being allowed to use it. Import what you have the right to transcribe — your own recordings, your own episodes, material you are licensed to work with, or content you are quoting within the limits of fair use where you are. The tool does not make that judgement for you.