Descript changed how many people edit audio and video. You upload a recording, it produces a transcript, and you cut, rearrange, and delete the media by editing the text like a document. Removing a sentence in the transcript removes it from the recording. For podcasts, interviews, tutorials, and talking-head videos, that is much faster than working on a waveform, and people who have never used an editor can pick it up quickly.
AI features sit around that core. Filler word removal strips out ums and long pauses, Studio Sound cleans up noise and improves a voice recorded in a poor room, and Eye Contact adjusts where the speaker appears to look. Underlord, the built-in assistant, can perform edits from instructions, write show notes, create clips for social media, and draft titles. Descript can also record remote guests, add captions, and use a synthetic copy of your own voice to fix a mistake without re-recording.
Descript has a free plan with limited features and export allowances, and paid tiers that add higher-quality export, more transcription hours, and AI credits. It is great for spoken content but less suited to music production or complex visual editing, where Audition, Premiere, or DaVinci Resolve go further. Riverside and Cleanvoice overlap on parts of the workflow. A common pattern is to rough-cut in Descript and finish elsewhere if needed.