How to edit video in Descript using text

How to edit video in Descript using text

Descript treats spoken video like a document. After it transcribes your recording, each word in the transcript is connected to the matching moment in the audio and video.

Delete a sentence from the script and Descript removes that section from the video. Cut and paste a passage to change the order, and the matching footage moves with it.

1. Create a project. Open Descript, start a video project and import an MP4, MOV or another supported media file. You can also record your screen, camera and microphone inside the app.
2. Choose the speaker language. Descript will transcribe the recording. Processing time depends on the recording length and your internet connection.
3. Check the transcript. Correct misspelled names, product terms and technical words before making large edits. Correcting text alone should not alter the spoken audio if you use the transcript correction option rather than replacing the media.
4. Remove unwanted lines. Highlight the words, sentence or paragraph you want to cut, then press Delete. Play across the new join and undo the change if it sounds rushed.
5. Reorder useful sections. Cut the selected transcript passage and paste it at a new point. Descript moves the connected media, so check that the picture still makes sense after the move.
6. Shorten pauses. Find long gaps in the transcript and reduce them rather than removing every pause. Small pauses help speech sound human and give viewers time to follow an idea.

How do you make clean cuts without jumpy video?

Text editing makes cutting fast, but it cannot hide every visual jump. If a speaker changes position between two sentences, deleting the words between them may create an obvious jump cut.

Start by listening to the join. Restore a few words or a short pause if the speech sounds clipped. A good edit usually matters more than making the video as short as possible.

Next, cover visible cuts with supporting footage, a screen recording, an image, a title card or a tighter crop. In Descript, place the extra visual over the cut while keeping the main spoken audio underneath.

Scenes help you change what appears on screen at different points in the script. Add a scene boundary where the visual should change, then apply a layout, background, camera crop or supporting visual to that scene.

Watch the picture and waveform around every major cut. Headphones make small clicks, repeated breaths and sudden room tone changes easier to hear.

How can you remove filler words and repair audio?

Descript can find common filler words such as “um” and “uh.” Open the editing or AI assistant controls, choose the filler word tool, then review the detected words before applying changes.

Removing every filler word can make a person sound unnatural. Keep fillers that show hesitation for a reason, preserve timing in a conversation or sit too close to another word for a clean cut.

Some plans may limit bulk filler word removal or other AI actions. Descript has a free plan, while paid access commonly costs around $15 to $30 per month at the time of writing, depending on billing and included AI usage. Check the current plan page before paying because limits and prices can change.

For background noise, echo and uneven speech, try Studio Sound. Apply it to the selected clip, start at a moderate intensity and compare the result with the original. Strong processing may make breaths or certain syllables sound artificial.

If you said the wrong word, the safest repair is often to record the sentence again. Put the microphone in a similar position, record in the same room and replace the full phrase rather than one tiny word.

Descript also offers AI speech correction features in eligible workflows. Use them only with your own voice or with permission, and listen for changes in tone, pronunciation and pacing. A clean rerecording is usually better when the correction sounds synthetic.

How do you add and format captions in Descript?

Captions make a video easier to follow with the sound off. Because Descript already has a transcript, you can turn that text into timed captions without typing it again.

1. Correct the transcript. Fix names, punctuation and misheard words first, since the captions use this text.
2. Add captions. Choose a caption style from the text or captions controls and apply it to the required scenes.
3. Set readable lines. Keep each caption short enough to scan quickly. Break long thoughts into smaller phrases instead of filling the screen with text.
4. Choose simple styling. Use a large font, strong contrast and a background or shadow when the footage is busy. Avoid decorative fonts for body captions.
5. Check placement. Keep text away from faces, app controls and the lower areas that social platforms may cover.
6. Review timing. Play the whole video and look for captions that arrive late, disappear early or split a name in an odd place.

You can burn captions into the exported picture when viewers need to see them everywhere. Depending on your project and plan, you may also be able to export a subtitle file for platforms that support optional captions.

For vertical social video, preview the project in the intended portrait layout. A caption that looks fine on a wide canvas can cover the speaker or run beyond a safe viewing area after resizing.

What export settings should beginners use?

Before exporting, watch the composition from beginning to end. Look for black frames, missing visuals, accidental text cuts, caption errors, volume jumps and unsupported media notices.

Open Descript’s publish or export controls and choose a local video file. MP4 is the practical choice for YouTube, social platforms, client review and most website uploads.

For a normal landscape video, use a 16:9 canvas and export at 1080p when available. For TikTok, Instagram Reels and YouTube Shorts, use a 9:16 portrait canvas and confirm that captions remain inside the visible area.

A higher resolution creates a larger file and cannot restore detail that was absent from the original recording. If your source is 1080p, a 1080p export is usually enough. Use higher settings only when the source footage and destination need them.

Give the exported file a useful name, such as the topic plus “final” and the date. Play the saved file outside Descript before uploading it, since this confirms that the export completed and that audio stays in sync.

FAQ