Charter Work Tech: The best AI podcast editors

Our process

Traditionally, podcast editing involved editing waveforms, but newer tools allow users to edit audio the same way they’d edit a document—by deleting words, copying and pasting sections, etc. We only included tools with text-based editors in our review, testing a total of four: Descript, Podcastle, Riverside, and Streamlabs.

We tested each tool's ability to edit a simple one-on-one conversation, evaluating them based on the suite of the editing tools they offer and their ability to make edits sound natural. Note that while all of these applications offer other features that are important to podcasters, such as the ability to conduct high-quality remote recordings, we only tested their ability to edit audio.

What to watch for: An increasing number of podcast editors are working to integrate text-based editing into their products. Riverside, for example, just rolled out the feature in March of this year.  Expect more options to choose from in the coming months.

Also keep an eye out for Adobe Podcast. The product is still in beta, but it also offers users the ability to text-edit audio. We didn’t get the chance to include Adobe Podcast in our review, but given the tech company’s track record of delivering great products, we expect the new tool to be a solid option for users.

Disclosure: In addition to our own testing, we were given a live demo of Descript. But given that Descript is so far ahead of the other products we tested, we don’t see this as an issue or skewing our evaluation.

Our recommendation

Descript stands alone as our top pick. Here’s how it performed on several key features:

Breadth of editing tools: On its website, Descript says that users can edit audio the way they’d edit a Google Doc. That’s a pretty accurate comparison, given all of the different editing options users have at their disposal. You can do simple tasks like delete filler words, rearrange a podcast by cutting and pasting sections, and edit tracks separately (many of the tools we tested don’t offer all of those features). You can also edit the underlying waveforms to smooth out changes, fade in and out, etc. And similar to Google Docs, Descript lets users collaborate on an audio project, allowing people to leave edits and comments for each other to see.

Something that really impressed us was Descript’s responsible use of its overdubbing feature. Imagine you’ve recorded a podcast with someone, and you realize afterward that they mumbled the end of a sentence. With the overdubbing feature, you can type in what they said and generate a clip of the original speaker saying it. Of course, such a feature could easily be abused.  To prevent users from overdubbing people without their permission, Descript requires a Voice ID statement from the speaker, explicitly giving their permission for Descript to recreate their voice.

Transcription accuracy: If you’re going to edit a podcast by editing the underlying transcript, it’s important for that transcript to be accurate. Descript’s transcription feature is powered by Rev, a top pick from Charter’s review of AI audio transcription tools. The transcript that Descript generated was very accurate, though it occasionally incorrectly transcribed words and attributed words and sentences to the wrong speaker.

Ease of use: Descript has a steeper learning curve than you might expect. It takes some time to learn about, locate, and get comfortable using all of its different features. It can also be a tad glitchy, at times resisting or undoing edits. Still, it’s a lot easier to use than other audio editors, like Pro Tools, and it doesn’t take long to get comfortable using it.

Supported languages: Descript can transcribe 23 different languages. Those working in a language not included should check out Riverside, which uses OpenAI’s speech recognition model Whisper and says it can support over 100 languages (note: Riverside has far fewer editing tools than Descript).

Pricing deep dive

Descript has four pricing options: Free, Creator, Pro, and Enterprise.

• The Free plan allows you to transcribe and edit up to one hour of audio per month. It lets you automatically remove the filler words “uh” and “um,” and it has a 1,001-word vocabulary for overdubbing.
• The Creator plan is $12 per user per month billed annually, or $15 per user per month billed monthly. Each user can transcribe and edit up to 10 hours of audio per month.
• The Pro plan is $24 per user per month billed annually, or $30 per user per month billed monthly. Each user can transcribe and edit up to 30 hours of audio per month. It can recognize and automatically cut up to 18 filler words and cut repeated words. It also has an unlimited vocabulary for overdubbing, meaning you can ask it to overdub anything. (You can read more about the overdubbing feature here.)
• The Enterprise plan includes everything that you get with Pro, plus extra features like a dedicated account representative, onboarding and training, and single sign-on. Inquire for pricing.

You can read more about the different pricing options here.

How we chose what to review

We started our search by reading several product reviews and blog posts, which surfaced many options to choose from. In order to winnow down our list, we only looked at applications that have text-based audio editing tools. Tools that require you to edit waveforms have higher barriers to entry, and they’re less practical for people who don’t already have experience editing audio. Over time, we expect to see a growing number of companies offering the text-based editing feature.

Sign up here to receive our free Work Tech newsletter in your inbox.