Introducing Descript
Writers have the ideal creative tool in the word processor. Learning to use it occupies a brief moment in a career otherwise dedicated to refining one’s craft. For print writers, tools rarely get in the way of one’s ability to express their ideas.
We’re living through a rapid shift in media consumption, from print to audio and video. Unlike the word processor, the creation tools for audio and video are complicated, slow, and require an ongoing diet of training.
Launching today with a $5M seed investment from Andreesen Horowitz, Descript aims to solve that problem, taking advantage of the recent explosion in speech recognition accuracy to press the reset button on how talk-driven media is created. We are making it faster, easier, and more fun.
What is Descript?
Descript is a toolkit of superpowers for anyone that works with voice audio. Here are the highlights:
1. Audio Word Processing
View voice audio files as text instead of waveforms — and then edit the audio by editing text.
2. Fast/Cheap/Accurate Transcription (by robots or humans, as needed)
Fast. Transcripts in under five minutes or we’ll buy you a pizza.
Cheap. Just 7¢ per minute of audio— pricing details here.
Accurate. Up to 95% accurate, with half as many errors as other transcription services. And if the state-of-the-art in automatic transcription isn’t enough…
Upgrade to White Glove. Automated transcription isn’t perfect. And while Descript makes it fast and easy to correct typos, sometimes you really can’t be bothered… especially when you’re dealing with low quality audio. For those situations, upgrade to White Glove transcription for 99¢ a minute, and humans will transcribe your audio in 24 hours or less.
3. Waveform Editor with a special trick
For fine-tuned trims and crossfades, Descript has a full-featured waveform editor. It’s got something you won’t find anywhere else — the Word Bar. Just grab a word tab to add/remove silence between words:
4. Web Publishing and Commenting
Publish your Descript composition to the Web, and share the link with friends and colleagues, who can listen and leave comments, Google Docs style. If you’ve ever had to give feedback on an audio file, you’ll love this a lot. See a live demo here.
5. Non-destructively export to other audio apps
After you’re done editing in Descript, you can export your session to a format that can be read by Pro Tools, Adobe Audition, and most other Digital Audio Workstations. That means it will preserve the original audio file, and all of your edits will be visible when you open it in the DAW.
How Descript became a thing
Descript began three years ago as a audio tour production tool for Detour. We wanted to make it possible for anyone to record and edit audio for Detour, but the popular audio production tools were designed primarily for music production, and had a high learning curve.
It occurred to us that it’d be easier and faster to work with audio if you could edit text instead of waveforms. We did some research, and found that automated transcription and text/audio alignment were reaching a technological tipping point where the concept was perhaps viable.
We quickly built a v1 and started showing it to audio producers, at which point we realized that just about everyone who has worked with audio has independently had the same idea, and indeed spent a portion of every day cursing its non-existence. Because, of course people would rather edit text than waveforms. The technology just didn’t exist.
Descript, Inc.
For the last two and a half years, we’ve been working arm-in-arm with audio producers to evolve Descript into their voice audio’s best friend. As the product matured, we realized that Descript shouldn’t remain subservient to Detour — it needed to chart its own course.
With this in mind, we decided to spin out Descript as a separate company, and raised $5M in seed financing from Andreesen Horowitz, with Alex Rampell joining our board. I got to know Alex when I was running Groupon and he was running TrialPay, and I’m excited to have the opportunity to work alongside such a smart and capable entrepreneur.
Last but not least, I’m thrilled that the one-of-a-kind Bing Gordon will be serving as an advisor to Descript.
What about Detour?
I‘ll have more Detour news in the near future, but for now, you’ll have to make due with this cryptic reassurance that good things lie ahead. As a full-time Descript employee, I’ll no longer be in an operating role at Detour, but there’s a super talented team manning the helm. I’m confident in the greatness of the hands it is in.
Where we’re headed
Today’s Descript — a single-track voice audio editor with automated transcription — is the tip of the iceberg. With a little imagination, you can guess where we’re going — a ground-up reimagining of how people create talk-driven media.
By shifting from timeline/waveform to script-based editing, Descript makes creation easier for newcomers, faster for professionals, and serves as the foundation for new forms of AI-powered media synthesis that are just beginning to emerge.
If that sounds like something you want to be part of, we’re hiring engineers to join our team of 8 in our San Francisco headquarters.
In closing, I’d like to bore the rest of you by publicly thanking the incredible team that’s been building Descript over the last three years. You boringly reading this paragraph is a beautiful way to show that you too are grateful for their hard work.
Download Descript and get your first 30 minutes of transcription for free.
