AWS Startups Blog

Could This New Speech-to-Text Tech Break the Internet?


Trint CEO Jeff Kofman on the Trint Player

Guest post by Michael Wolter, Content Marketing Manager at Trint

You might remember when Kim Kardashian’s 2014 Paper magazine cover boldly stated it would “Break the Internet.” All due apologies to Ms. Kardashian, but the Internet didn’t actually break as a result of her racy cover—it went on functioning normally.

The truth is the Internet can’t be broken. It doesn’t work like that. Bad actors can try to cause regional disruptions, but aside from a cataclysmic event like the meteor that finished off the dinosaurs, the Internet will keep plugging along.

But the internet can be slowed down, and one of the ways that can happen is an influx of data—a whole lot of data.

At audio and video transcription company Trint, this is an issue that we think about a lot, especially as we gear up to launch our latest feature, the Trint Player. The Trint Player turns our interactive transcript, a productivity tool for media production, into an interactive video player for consumers that can be posted onto any webpage. The result is a little bit like karaoke: each word is highlighted as it’s spoken in the video.

The Trint Player also gives online video producers a new way to make the most of their video content. Video producers can post stories online, which can then be watched without sound—that means that even if you’re watching on a bus, at work, or in the back row of a wedding, you can still keep track of what’s going on in the video. Marketing departments can make their online video searchable in Google by attaching the interactive transcript. And anyone with accessibility issues, like people who are deaf, can access video content where previously they couldn’t.

But we want our technology to get bigger than that. If the uptake is fast enough, the Trint Player could introduce a monumental amount of new data to the world wide web in a very short amount of time. Online video is getting very popular, very quickly: by 2021, 1 million minutes (about 17,000 hours) of video content will cross global IP networks each second. If the Trint Player embedded even 1% of the new video added, the amount of new data on the Internet would come hard, fast and relentless—like a tsunami that keeps on crashing.

So, is this impending data something we need to worry about? Should we a bunker and stock it with canned food and bottled water?

Simon Turvey, VP of Engineering at Trint, says we don’t need to worry. “We built the Trint Player to be very lightweight,” he says. “When it comes to the video in the Trint Player, instead of replicating existing video content hosting we seamlessly allow customers to leverage their existing hosting solutions, like YouTube. This means we’re not duplicating video content that’s already on the Internet, just playing it from a different source.”

“As for the amount of data in the interactive transcript, text takes up very, very little data—500 pages of text equals about one megabyte,” he says. “The top four storage and service companies currently have about 1,200 petabytes among them. That’s enough to store 600,000,000,000,000 (600 trillion) pages of text.”

Not even Sally Jessy Raphael could talk long enough to make a dent in that.

So the Trint Player may not break the Internet, but it will certainly change the way video is produced and distributed. Video can currently only be searched by the words in the title and description, locking the spoken words in the video until it’s played. This is dark data. The Trint Player shines light on that data, allowing the content to be published and searched online.

Speech-to-text won’t break the Internet, but it does promise to change how online video is published. Interested? Email to learn more.