Generate AI images in Descript

Demo summary
The user types a text prompt for a 'chicken salad' into Descript's Generate Image tool to create a custom visual asset for the video.
Step-by-step
- Open the media library or archive menu
- Select the Generate Image feature
- Type a descriptive prompt into the text field
- Wait for the image to generate
- Overlay the generated image onto your video timeline
Tips
- Use specific descriptions like 'clean, minimal worktop' to get more accurate visual results
- Use generated images to emphasize specific points while you are speaking
Highlights
“Wait just a few seconds and voila, we now have exactly what we were looking for”
All demos from “How to EDIT Videos 10x FASTER Using AI (Descript Overview)”
0:430:16Text-based video editing in DescriptThe user demonstrates how deleting a sentence from the automatically generated transcript in Descript simultaneously removes the corresponding footage from the video timeline.DDescript· AI Video Editor
1:130:15Remove filler words with UnderlordThe video shows the Underlord AI assistant in Descript automatically detecting and deleting filler words like 'ums' and 'likes' from a talking head video.DDescript· AI Video Editor
2:310:43Adding B-roll and sound effects from Descript libraryThe creator searches and inserts stock waterfall footage and cinematic sound effects directly from Descript's built-in media archive.DDescript· AI Video Editor
3:270:28Generate AI images in DescriptCurrentThe user types a text prompt for a 'chicken salad' into Descript's Generate Image tool to create a custom visual asset for the video.DDescript· AI Video Editor- Watch “How to EDIT Videos 10x FASTER Using AI (Descript Overview)” →
AI Video Editor
0:430:16Text-based video editing in DescriptThe user demonstrates how deleting a sentence from the automatically generated transcript in Descript simultaneously removes the corresponding footage from the video timeline.Learn Online Video
1:130:15Remove filler words with UnderlordThe video shows the Underlord AI assistant in Descript automatically detecting and deleting filler words like 'ums' and 'likes' from a talking head video.Learn Online Video
2:310:43Adding B-roll and sound effects from Descript libraryThe creator searches and inserts stock waterfall footage and cinematic sound effects directly from Descript's built-in media archive.Learn Online Video
3:270:28Generate AI images in DescriptCurrentThe user types a text prompt for a 'chicken salad' into Descript's Generate Image tool to create a custom visual asset for the video.Learn Online Video
2:261:10Text-based video editing in DescriptThe creator demonstrates how to upload a video and edit it by highlighting and deleting text from the automatically generated transcript, which removes the corresponding video footage.Vince Opra
4:330:36Using the Blade tool on the timelineThe video shows how to use the Blade tool (shortcut B) to manually split clips on the timeline and delete specific segments.Vince Opra
7:150:46Remove filler words with Underlord AIThe demonstration shows Descript's Underlord AI identifying and batch-deleting filler words like 'um' and 'uh' from a recording.Vince Opra
10:570:32Add automated captions to videoThe demonstration shows how to apply and customize dynamic captions to a video project using Descript's built-in captioning styles.Vince Opra
0:310:40Correct eye contact with Descript AIThe creator demonstrates the 'Eye Contact' feature in Descript, which digitally adjusts his eyes to look at the camera lens even while he is reading a script off-screen.Done By Lunch
1:110:25Shorten word gaps in DescriptThe user shows how to use the 'Shorten word gaps' tool to automatically identify and remove silences longer than one second across a 14-minute video file.Done By Lunch
1:360:40Remove retakes automatically in DescriptThe 'Remove retake' feature is shown identifying repeated phrases in a transcript and automatically selecting the best take to keep while ignoring the others.Done By Lunch
2:160:34Automated multicam editing in DescriptThe creator demonstrates 'Automatic Multicam' which processes two separate video files and automatically cuts between speakers to create a finished interview layout.Done By Lunch