txscribe

Transcripts

Public recordings, transcribed

Press briefings, a Supreme Court argument, open meetings, public lectures and interviews — run through the same speech recognition and summaries as your own recording, and left as the machine wrote them. Every transcript plays against the publisher's original audio: click a paragraph and hear whether it says what was said.

By publisher

All recordings

How these were made

Each recording went through the same speech recognition and the same summaries every txscribe user gets. Nobody has corrected the result. Where a transcript mishears a name or a number, that is what the product would have given you too — which is why each page plays the original audio beside it.

Speakers are numbered in the order they first speak, as they are for anyone's upload. The names you see on a page are the ones the publisher lists, and which number belongs to which name is left to you, the same as in the app.

Only recordings anyone may republish are here: work of the US federal government, programmes Voice of America produced itself, and interviews their authors released under Creative Commons licences that allow it. Each page names its licence and links to the publisher's own copy. That is why you won't find last night's podcast.