Descript

What it is

Descript edits video and audio by editing text. It transcribes your recording, and when you delete a sentence in the transcript, that sentence disappears from the video. For anyone who has scrubbed a timeline hunting for the moment someone said "um", the change of interface is the whole product.

The AI tools sit on top of that idea. Remove Filler Words strips verbal tics in one pass, Studio Sound cleans up audio recorded in a normal room, Eye Contact adjusts where you appear to be looking, Green Screen removes a background without a physical one, and Underlord acts as a co-editor. It transcribes in more than 25 languages and can translate and dub into more than 30.

Who it's for (and who it isn't)

It fits podcasters, course creators, marketers and anyone producing talking-head video regularly who is not a trained editor. The transcript-first approach means the skill you need is reading and cutting text, which most people already have.

It is a weaker fit for cinematic or heavily designed video work, where a conventional timeline editor gives you control Descript does not try to offer. It is also not the tool for a one-off video: the free plan gives you 60 minutes of transcription per month and exports carry a watermark at 720p, so a single project either fits in that or needs a paid month.

Pricing

Descript is freemium. The free plan includes 60 minutes of transcription per month, 100 one-time AI credits, and watermarked exports at 720p with limited access to the premium AI features. Hobbyist is $16 per user per month billed annually, or $24 monthly. Creator is $24 annually or $35 monthly. Business is $50 annually or $65 monthly, and Enterprise is custom. Higher tiers raise the monthly AI credits, from 400 up to 1,500, increase processing hours from 10 to 40, and add 4K and 5K export.

Two limits decide your plan: transcription hours per month and AI credits. Estimate the hours of raw footage you record, not the length of the finished video, because you transcribe everything you shoot.

Set it up

Official guide: Get started with Descript

1. Have one short recording of your own ready, a video or audio file, because transcription uses up the Free plan's 60 minutes per month. Create a Descript account on the Free plan, which also has 100 one-time AI credits (the allowance Descript's AI features draw from) and watermarked 720p exports. Your Free account is now ready.
2. In Descript, choose "Import" and drag in your recording. Descript transcribes all media for you, which means it turns the speech into text. Then open the "Project" panel in the right sidebar and find the recording under "Files". Drag it into the script editor. You now see your words as text.
3. Editing is non-destructive, so you can bring anything back: right-click the edit boundary, the spot where text was cut, and select "Restore removed media". Now select a sentence you want gone and press "Delete" or "Backspace". The text is now gone from the transcript, and the matching audio and video are gone from your composition.
4. Open the "AI Tools" panel in the right sidebar and choose "Remove filler words". Filler words are the "um" and "uh" in speech, and Descript underlines them in light blue. This works for English recordings only and uses AI credits. For each word you want gone, pick "Delete". Your chosen filler words are now gone from the text and the audio.
5. Select your script layer, the part of the project that holds your recording, and open the "Properties" panel in the right sidebar. In the "Audio Effects" section, toggle on "Studio Sound". It cleans up the sound and needs an internet connection. To make the effect weaker, lower the "Intensity" slider. Studio Sound is now on for the whole file.
6. Click "Export" in the upper-right corner and select the "Video" tab. Choose "Local export" as the destination, adjust the settings and click "Export". On the Free plan, the video has a Descript watermark in the bottom-right corner. Your video is now saved to your computer as an MP4 file.

When this works, you have turned one real recording into a cleaner video file by editing text, and you can see whether the Free plan limits are acceptable for your work.

On Hobbyist ($16/user/mo billed annually ($24 monthly)): Studio Sound and Remove filler words are no longer limited, and exports carry no watermark. Exports you already made stay watermarked until you export them again.

Strengths and limitations

The strength is speed for talking-based content. Transcript editing collapses the slowest part of podcast and interview production, and the AI cleanup tools remove the second slowest.

The limitations are real and worth planning around. Transcription accuracy depends on audio quality and accent, and a bad transcript makes the whole method frustrating. The free plan's 60 minutes and watermarked 720p exports are a demonstration rather than a working tier. AI credits meter the features people actually came for. And for anything requiring precise frame-level control, this is the wrong interface, not a cheaper one.