Descript

Descript is an editor for podcasts and video that works off a transcript: you edit the words in a document and the audio and video follow. That approach makes cutting, rearranging, and cleaning up recordings much faster than dragging clips on a traditional timeline.

The product has shifted decisively toward AI. Its headline layer is now Underlord, a generative co-editor that drafts scripts, writes show notes, automatically surfaces viral and social clips, reorganizes scenes, and can generate and extend video on its own. Classic tools like filler-word removal, noise suppression, and timing fixes are still there but are now framed as part of this AI workflow. Descript has also added an AI model picker that lets you choose between models from Anthropic, Google, and OpenAI (or an Auto option) to control reasoning and style, parallel AI media-generation jobs, automatic avoidance of letterbox and crop bars, and an API, currently in beta and open to all users, with MCP support.

On the editing side you get multitrack audio and video, screen recording, layered timelines, remote recording, transcription, and AI voices, plus AI Speech-based corrections and translation and dubbing features.

Pricing is freemium. There is a Free plan at $0 (60 media minutes per month and 100 one-time AI credits, with limited AI tools), Hobbyist at $16/user/month billed annually ($24 monthly), Creator at $24/user/month annually ($35 monthly), Business at $50/user/month annually ($65 monthly), and a custom Enterprise tier. The billing model has been rebuilt so that the old transcription-hours allowance was replaced by monthly media minutes/hours plus metered AI credits, with top-ups available.

The main limitations are cost predictability and dependence on credits. Because AI generation and media time are metered, heavy AI users can burn through their allowance and need top-ups, and the plan structure is more complex than a flat subscription. It is also a fairly heavy app that expects reasonably capable hardware.

Descript suits podcasters, video creators, marketers, and teams who want to edit by editing text and lean on AI for drafting, clipping, and cleanup rather than hand-crafting every cut on a timeline.