Repurpose Client Content into Clips with Descript

1. Go to descript.com and click Sign up
2. Use your work email or sign in with Google
3. You'll land on a project dashboard once verified

What you should see: An empty project list with a button to start a new project. Troubleshooting: If verification email doesn't arrive, check spam before requesting a new link.

Before you upload anything, this step matters more than any button in the app.

1. Check the client's contract or engagement terms for language covering content ownership and repurposing rights
2. If the recording includes the client speaking, confirm they consented to being recorded and to having clips shared publicly
3. In states requiring two-party consent for recordings, confirm that consent was captured at the time of recording, not assumed after the fact

What you should see: Nothing in the app yet, this is a check against your contract and notes, not a Descript screen. Troubleshooting: If you're not sure the client agreed to public repurposing, ask before you publish anything, even if the underlying call recording itself was already consented to.

1. Click New Project, then upload your video or audio file
2. Descript automatically transcribes the full recording once uploaded
3. Wait for transcription to finish; longer recordings take a few minutes

What you should see: A text transcript appear alongside the video player, with each spoken word timed to the footage. Troubleshooting: If a name or technical term gets transcribed wrong, this is normal for auto-transcription; you'll fix it during editing in the next step.

1. Look for Underlord in the toolbar, Descript's AI agent
2. Ask it to identify strong short clips from the recording, or use a built-in "Find clips" style feature if shown
3. Review the suggested moments rather than accepting them automatically

What you should see: A list of suggested time ranges, often with a short description of why each was flagged (a strong quote, a clear insight, a complete thought). Troubleshooting: If none of the suggestions feel usable, scan the transcript manually for a self-contained 30 to 60 second moment; Underlord works better on longer, more free-flowing recordings than tightly scripted ones.

1. Select the clip range you want to keep
2. Delete filler words, false starts, or tangents directly in the transcript text. Removing the text removes it from the video
3. Add captions using the built-in caption tool if they're not already generated

What you should see: The video preview update in real time as you delete text, with the corresponding footage cut automatically. Troubleshooting: If a cut feels abrupt, add a small pause back by leaving a beat of silence in the transcript rather than cutting flush against the next word.

1. Once a clip is finalized, use Export
2. Choose the format and aspect ratio for where it's going: vertical for Reels or TikTok, square or horizontal for other placements
3. Download or publish depending on your workflow

What you should see: An export queue showing progress, then a downloadable file or direct publish option depending on your account's connections. Troubleshooting: If a vertical export crops out something important (a slide, a second speaker), reframe the clip manually before exporting rather than trusting the automatic crop.