Descript harnesses the power of AI for easy audio and video editing

A few years ago, Descript burst onto the scene as an entirely new approach to editing audio. For anyone working on podcasts, it’s a vastly different way of doing things compared to traditionally complex audio tools. 

How Descript works is simple. First, it automatically transcribes all your uploaded audio (thanks to artificial intelligence), then you edit your recordings by simply highlighting and deleting (or moving) any words or passages in a text editor.

Earlier this year, I started a new podcast and picked Descript for all my editing. At first, I wasn’t sure it was robust enough. A dozen episodes later, I’m convinced their approach is a much faster and easier way to edit audio.

If you’ve never tried it, you’re in luck: Descript recently revamped its desktop app, adding video storyboarding, editing, and production tools in addition to existing audio functionality. Descript does require a paid subscription for anything more than testing, but plans are reasonable and priced on the low end for casual users, as well as offering robust plans for more professional production needs.

Let’s take a quick tour to see it in action.

How to use Descript: A podcast editing walkthrough

I do a podcast together with my friend Joel even though we live thousands of miles apart. We record our conversation together, then I edit the sound files to make the show. Above is a screenshot of Descript’s desktop app after I imported two .wav file recordings into it, one being my voice (the purple one that starts with Matt) and the other being my co-host. 

Once the files are imported and transcribed, the first thing I do is run the Remove filler words feature in Descript. It scans your transcript to highlight hundreds of instances of “um” and “ah” that can be removed instantly with a single click.

Next, I press play and listen through the whole podcast while reading the transcript. When I spot any moments like one in the first screenshot below—as my co-host was completing a thought—I stop playback, highlight the repeated words, and hit delete. (If you’re on the Pro plan, even this part is automatic.)

Keep in mind Descript isn’t just changing the transcript: the audio files also update instantly to reflect it. After a quick edit, I re-listen, and the section not only sounds more natural, but Descript also smooths out your audio so no one will hear an abrupt cut. 

After over a decade of my own podcast production experience, I can say that 90% of the work in podcast editing is doing this, which Descript makes much faster and easier. Using professional audio tools, I’d be cutting and splicing waveforms for hours by hand to achieve similar results.

Descript hides complexity from the user, but it also lets you go deep when needed. If one of your hosts accidentally interrupts the other in a recording (audio folks call this “cross-talk”), you can highlight the exchange and use the Edit sequence feature to get a more precise editing interface. Here, I simply select the interrupting speaker’s segment I want to remove, then hit delete and click Done to save.

I continue the process of listening and cutting until I get through the whole episode, which usually takes less than twice as long as the recording time. Then I export out the final audio file to upload to my podcast publishing system. In the past, using professional-style audio apps, I would typically spend 4-10x the recording time adjusting waveforms, cutting out gaps and filler words, redoing levels on audio, and all the other tedious work—everything that’s automated by Descript.

Magical audio powers you’ll only find in Descript

Descript uses AI to cut out tons of busywork, add entirely new capabilities not seen elsewhere, and save editors tons of time.

One of the most mind-blowing Descript AI features is Overdub. It lets you insert words to your transcript, which are added into your recording. It’ll use one of their stock AI voices (one amusing example sounds like the deep-voiced narrator you often hear in movie previews), or you can even train the AI to reproduce your own voice.

Ever record a great conversation with a guest but you forgot to mention one of their recent credits in your introduction? With Descript Overdub, you can add a few words here and there that you forgot to say during your live recording, and it will generate your own voice with startling accuracy, embedded into your recording exactly where you place it. I’ve tested it myself, and it’s really remarkable stuff you can’t find elsewhere.

Descript uses AI for other useful features in the app too. Some examples:

• The transcripts that power it take just a few minutes to produce, even when you upload over an hour of audio. 

• The identification and immediate removal of filler words saves additional editing time. 

• You can remove audio gaps between speakers to speed up your show and remove any dreaded dead air. 

• When you’re ready to publish, the final audio files created are auto-spliced and leveled to make all your speakers sound like they were recorded together in the same room. 

Video production comes to Descript

Descript recently added video features to their suite of editing tools. Take a look at this overview video they made to see all the options.

Similar to their approach to audio editing (simply manipulating text on a page instead of being mired in a complex software interface), Descript video editing makes putting clips together for an informative video more like making a PowerPoint presentation than editing a film.

Here, I’m working on a review of a truck I bought a few months ago. I created a new video project in Descript, gave it a title, then started storyboarding by creating some scenes, each with a quick summary. For the introductory scene, I wrote out a paragraph of what I wanted to say, then I hit record and captured a new sound file as my narration for the video. I went out and shot a few more clips, associating each of them with a scene. Once I finish each scene’s script, I’ll record audio for those, then export a final version of the video for YouTube.

The video features are new, but they show a lot of promise, again offering a much more text-first and human-friendly way to craft your stories in a text editor instead of a complex interface covered in waveforms and windows with dozens of editing options. 

Connect Descript to your other apps

Descript software automates a lot of the audio and video editing process, and you can take that automation one step further by connecting it to the other apps you use.

With Zapier’s Descript integrations, you can automate things like having Descript import new audio files and transcribe them the moment they finish uploading into a cloud storage account. 

It works the other way around too, saving your transcripts as files:

Or you can fire off a message or email when new transcripts are ready—to let people know editing can begin or to alert a team that the production process has started. 

Streamline audio and video production with Descript

Descript really excels at bringing all the power found in professional audio and video editing apps to anyone that can use a word processor. I’ve spent more than ten years using apps like GarageBand, Audacity, iMovie, Final Cut, and Premiere, and whenever I learn a friend works on podcast production, I always mention Descript to them as a better option. So if you’ve ever thought of starting a podcast, or you want to share your work on YouTube or TikTok, Descript is a great place to start.

Related reading:

• The best AI video generators

• The best AI productivity tools

• The best audio editing software

• The best transcription services

• What is Otter.ai?