they should do an integration with Krisp. alternatively, you could record w/ zoom using krisp and it will remove it prior to import

Descript isn’t yet part of my recording pipeline because it’s missing this feature. I essentially want the same service Alitu is offering: https://alitu.com/

Happy to pay the same price; I just want it integrated in the Descript workflow.

Reduction is good. Elimination is better. I'd like to be able to give Descript a sample of breaths, for example, and have it find all the examples close to it - give or take, say, 20% or something - and highlight them. Then I could individually choose to either reduce them by some chosen percentage or blow them out altogether, without cutting off the bottom of all the other sound. Just the lump. Same thing goes for sniffs, burps, rattles, lip smacks and especially slurps and swallows. Teach Descript what it looks like, have it hunt down those and ones similar to it, decrease or axe. Done. Then shorten up the silences left behind.

While a nice feature, noise reduction is incredibly hard to do well, and would likely require a lot of time/resources to build something not as good as what iZotope RX offers. If this is a feature you tackle, I'd look at licensing from iZotope like Adobe has done with Audition rather than building something from scratch. Alternatively, adding the ability to use Audio Units or VSTs to bring in tools like those for people using Descript as their full production app would allow people to use any 3rd-party noise canceling tools.

please I don't want to have to use auphonic again

If I could get away from Auphonic I would move my workflow to Descript instantly.

+1 to this.

Specifically: Many podcast interviews are 2 tracks of 2 people taking turns speaking.

While Person A is speaking, Person B is not, BUT their audio is still picking up swallowing noises, mic scratches, breathing and other background noise. I call this "white noise" and remove it manually with Garageband BEFORE I upload file to Descript.

If there were a way to detect and remove the white noise, that would eliminate a major pre-processing step for me.

Thank you!

+1. This, plus the ability to add 3rd party VSTs, etc, would be incredible.

YES please! Lots of new podcasters are using iPhones and their earbuds which could use some good canceling.

Event a simple noise gate would suffice. You could use the method that's open source in Audacity.