twaud.io

» see twaud.io as it lived, on the wayback machine

~ the goodbye ~

For about four years Twitter could not carry a file. You got 140 characters and a link, and if you wanted to show somebody a photo you used Twitpic, because Twitter had no idea what a photo was. An entire economy of small companies grew in that gap, each one taking a single file type and building a short link for it.

twaud.io took audio. It launched in May 2009 out of Massive Robot, a London studio run by the British developer Dan Webb, and its own homepage described it in exactly the terms everybody was using: think of it as Twitpic but for mp3s. You uploaded a file or recorded straight into the page, and if you were away from a computer you could email the clip in, which worked nicely with iPhone Voice Memos. The site then tweeted on your behalf with a short link to a page where your followers could listen and leave comments, and it told you how many people had shown up. There was a 10 MB ceiling and an API, and the footer carried a line insisting that twaud.io was not affiliated with Twitter inc.

What went through it was not what you would guess. The last archived front page is a list of recently posted clips, and it runs through a Spanish-language political recording, Japanese cover songs, a ringtone made for an Argentinian radio bit, and someone's take on a Selena Gomez single. Nobody had planned for that. They had built a link shortener for sound and the whole world had turned up with whatever it had on its phone.

~ the chronicle ~

~ the notice, and the part after the notice ~

On 11 June 2011 the top of the homepage turned into a grey band with an apology in it. twaud.io would be closing down. Record and upload were disabled as of that moment, but for the next couple of months you could still log in and download your own audio. SoundCloud was recommended as the alternative, with a mention of its iPhone and Android apps.

That promise did not survive. Archive Team's note on the shutdown records that the download and playback functions were switched off with no further warning, well before the couple of months were up. The consequence was immediate and total: every twaud.io link that had ever been published on Twitter stopped working on the same day. That is thousands of tweets whose entire content was a pointer at a file nobody could reach any more.

~ what the crawlers found afterward ~

The site itself did not vanish on the 11th. The homepage kept answering into September 2011, shrinking as it went, from 3,671 bytes in June down to about 2,370 by the start of the autumn. The marketing copy was still sitting there under the apology, promising that this was just the start of what twaud.io would offer and inviting you to follow the account for news about new developments.

By May 2013 the domain was serving a Gandi parking page in French and English, telling visitors the name was registered and currently parked by its owner. It still 301s today, fifteen years after launch.

~ the salvage ~

Archive Team got there during the shutdown window and built a rescue script around the one thing still reachable: the last twenty clips from each user profile. Volunteers ran it against a central tracker that handed out usernames, with a Windows version written in JScript for people who could not run the shell script.

What came out of that is on archive.org as a collection covering 2009 to 2011, somewhere around 21,000 mp3s from roughly 4,300 accounts. Webb's other credit in those years was Twitter's own platform team, which is a tidy place for the story to sit, given that the platform's gaps are what made twaud.io necessary at all.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »