Table of Contents
What Descript Actually Does
Descript transcribes your audio or video, then lets you edit the recording by editing the text — delete a sentence from the transcript, and the corresponding audio/video is cut automatically. Beyond editing, it adds Overdub (AI voice cloning for fixing errors without re-recording), Studio Sound (AI noise removal), and Storyboard (a simplified video editor).
Editing by deleting text instead of scrubbing a timeline
Cutting filler words, rearranging segments, and removing mistakes is dramatically faster when you're editing readable text rather than hunting for waveform boundaries — this is the core time savings over a traditional DAW or video editor.
Getting Started: Your First Project
Go to descript.com and sign up for the free plan — 1 hour of transcription per month, enough to test the workflow before paying.
- Click New Project and upload an audio or video file (or record directly inside Descript)
- Descript auto-transcribes the file — accuracy is typically 95-97% for clear English speech
- Once transcription finishes, you land in the text editor with your audio/video synced to the transcript
Text-Based Editing
How to Cut Content by Editing Text
- Read through the transcript and highlight any text you want removed — filler words, false starts, off-topic tangents
- Press delete — the corresponding audio/video segment is removed and the recording auto-splices together
- Use Filler Word Removal (under the Edit menu) to auto-detect and remove "um," "uh," and similar filler across the entire transcript in one pass
- Play back to confirm cuts sound natural — occasionally a manual adjustment is needed at a cut point
Studio Sound: AI Noise Removal
How to Apply Studio Sound
- Select your audio track (or the full project)
- Click Studio Sound in the effects panel
- Descript applies AI noise removal and presence boost automatically — no manual EQ or gate settings needed
In testing, Studio Sound reduced background noise by roughly 80% on a moderately noisy recording — not as precise as manual work in a dedicated tool like iZotope RX, but dramatically faster for typical podcast-quality cleanup.
Overdub: Voice Cloning for Fixes
Overdub lets you type new words and have them rendered in your own cloned voice — built specifically for fixing errors, not full content generation.
How to Set Up and Use Overdub
- Go to Overdub settings and record 30+ minutes of clean voice samples to train your clone (required before first use)
- Once trained, highlight a mispronounced word or a section you want to change in the transcript
- Type the replacement text
- Descript generates the new audio in your cloned voice and splices it into the timeline automatically
Use Overdub for short fixes, not long new sentences
In testing, short word or sentence-level fixes were nearly seamless — longer, more complex generated sentences showed slight artifacts. Reserve Overdub for genuine error correction, not writing new content from scratch.
Storyboard: Video Editing
How to Edit Video with Storyboard
- Open a video project — Storyboard shows your transcript alongside a scene-by-scene thumbnail view
- Rearrange scenes by dragging them in the Storyboard view, same as editing a slide deck
- Add B-roll, text overlays, and transitions from the panel on the right
- Export when finished
Storyboard is basic compared to Adobe Premiere but significantly faster for straightforward talking-head or tutorial-style video content where the main editing task is trimming and rearranging, not complex effects work.
Full Workflow: Editing a Podcast Episode
Here's the workflow that produced the ~40% editing-time reduction in testing:
Upload and let Descript transcribe
Transcription runs automatically on upload — this replaces the manual "listen through once" step of traditional editing.
Run Filler Word Removal, then read through for content cuts
Auto-remove fillers first, then manually cut tangents and mistakes by deleting transcript text.
Apply Studio Sound
One click cleans up background noise across the full episode.
Fix any mispronunciations with Overdub
Rather than re-recording a segment, type the correction and let Overdub generate it in your voice.
Descript Pricing 2026
| Plan | Price | Includes |
|---|---|---|
| Free | $0 | 1 hour transcription/month |
| Hobbyist | $24/mo | More transcription, Overdub |
| Creator | $40/mo | Full features for professional creators |
Start free — 1 hour of transcription is enough to test the full text-editing workflow on a real recording.
Common Beginner Mistakes
- Skipping Filler Word Removal. Running it first before manual editing saves real time versus removing fillers one by one.
- Expecting Overdub to write new content. It's built for short error fixes, not generating new dialogue — longer generated sentences show more artifacts.
- Using Storyboard for complex video work. It's fast for talking-head and tutorial content, but not a Premiere replacement for effects-heavy editing.