Turn long videos into short clips
Find the highlights in a long video, such as a YouTube video, a lecture, or a live stream, and cut them into several vertical short clips.
Clips cuts a long video into several vertical clips that keep the original picture and voice. Let AI find the highlights, or pick the parts yourself. To make a new video from scratch, see Create Reels and Shorts.
Create clips
Open the create panel
Open Videos in the sidebar and click Create clips, or choose New post → Create clips at the top of the screen.
Add the original video
On the Link tab, paste a YouTube link and click Import. Or, on the File tab, upload a video file. You can upload files up to 3 hours and 10 GB, import them from Google Drive, or pick a video you uploaded before.
After the video is checked, you see its length and the Estimated cost.
Set up the clips
Click Set up clips and choose a mode.
- AI mode: AI finds the highlights and creates clip candidates.
- Manual mode: you pick the parts to cut yourself, without AI.
If you need to, set these options too:
- Video template: a design for the title, captions, and layout. You only see templates with the same shape as your video (landscape, square, or portrait). Some templates, such as Stacked frames clip, show two shots one above the other.
- Target range: only the part you pick is imported and analyzed. The estimated cost is updated for that length.
- Clip length: leave it on Auto or pick a length. AI mode only.
- Playback speed: speeds up the video, voice, and captions. The pitch of the voice stays the same.
- Subtitle file: add an SRT or VTT file to use its subtitles instead of automatic transcription.
- Additional request: describe the scenes you want, such as "Pick only the funny moments." AI picks scenes from what people say, so it can't find scenes without speech, such as sports plays or scenery. AI mode only.
Start the analysis
Check the box to confirm you have the rights to the video, then click Find highlights. In manual mode, click Cut manually. The analysis continues in the video jobs panel at the bottom right of the screen. A 10-minute video usually finishes within 5 minutes.
Pick and create clips
When the analysis is done, open the project from the video jobs panel or the Clips tab. Pick a clip candidate in the list on the left, preview it, and click Create as is. You can also select several candidates and create them at once.
If AI missed a scene, go to that scene in Full video in the list and click Add range to add it yourself. In manual mode, you add the parts to cut the same way. If no scene matches your additional request, the project opens without candidates, and we refund the analysis credits.
Polish a clip
When you pick a clip, you can change its settings on the right. Your changes are saved automatically. The options you see depend on the template.
- Title and Subheading: edit the text AI wrote, or turn off Show on video to remove it from the video.
- Range: move the start and end of the clip. You can also use Set start here and Set end here to match the playback position. When you release a handle, it snaps to a sentence boundary within 2 seconds, so the clip does not start or end in the middle of speech. To keep the exact position, hold Alt (⌥) while you drag.
- Join another part: joins separate parts of the original video into one clip. You can join up to 5 parts, and they play in their original order.
- Remove silences: cuts the parts where nobody speaks.
- Remove original subtitles: finds and removes subtitles burned into the bottom of the original video. This uses more credits and takes longer to create.
- Playback speed: changes the speed of this clip only.
- Captions: turn captions on or off. For clips without a template, you can also pick a style, such as Word pop or Phrase.
- Highlight hook: AI picks an eye-catching moment from the clip and shows it first.
- Comment card: puts a card that looks like a viewer comment on the video. AI writes the comments, and you can edit them.
Use the Safe zones button at the left end of the button row under the preview to see where YouTube, Instagram, or TikTok buttons and captions cover the video. Check that the title and captions stay inside the dashed box. It shows only in the preview and isn't added to the video. For details, see Check safe zones.
Name spellings and forbidden phrases
In a workspace with brand rules, the rules apply to clips too.
- The titles, subheadings, and comment cards that AI writes use the official spellings and avoid forbidden phrases.
- If the transcription gets a name wrong in the captions, add that spelling to Commonly mistaken spellings in the rules, and the captions show the official spelling. When you change the rules, only the clips whose captions change are marked "Edited". Click Create again to apply the change.
- If a forbidden phrase is in the title, subheading, captions, or comment cards of the clip you're watching, a notice appears at the top of the screen. Click a title, subheading, or comment card line to go to that field in the settings. To fix captions, click Edit and change them in the editor.
Follow the speaker
When you pick a template that puts a landscape video into a portrait frame, Posteady finds who is speaking in each clip and keeps the frame on that person. It turns on by itself, so you don't need to set anything. While it searches, you see "Finding who is speaking" below the preview, and creating continues when it's done. If no speaker is found, the clip uses the position set by the template, and you can click Try again to search once more. Clips that follow the speaker use a few more credits when you create them.
Turn it off or on for each clip: if the preview shows the wrong part of the picture, turn off Follow the speaker in the settings on the right. That clip then uses the position set by the template and doesn't use the extra credits. If you turn it on again, Posteady reuses what it already found.
Scenes it doesn't follow well: in the scenes below, the picture can differ from what you expect. Turn it off for that clip, or pick a landscape-frame template that matches the original video.
- The speaker is off screen: when the camera shows the audience while the presenter talks, the frame stays on the largest face on screen.
- A person appears in a small window next to slides: following that person cuts off the text on the slides.
- A wide shot with people sitting far apart: zoom has a limit, so faces can look small.
- Several people laugh or talk at once: the frame can land on the wrong person. Very short interruptions aren't followed.
Fix only some scenes: if only a few scenes are wrong, keep it on and click Edit to move the clip to the editor. The video is split into pieces wherever the picture changes. In Timeline, select the wrong piece, then use Focal point under Media in Properties on the right. Click or drag on the video to move the part that shows. The zoom level stays the same.
Check where subtitles were removed
Clips created with Remove original subtitles on get a Show original button below the rendered video. Go to the scene you want to check and hold the button to see the same scene of the original, before its subtitles were removed. Release the button to return to the rendered video. Switch between the two to check that the area is clean. You can check videos rendered in the editor the same way.
The original is shown in preview quality, so it can look less sharp than the rendered video. The button doesn't appear on a video you haven't created again after changing its settings or editing it in the editor. It appears when you create the video again.
Polish a clip on your phone
On a phone, only the video stays on the screen. You use the buttons at the bottom of the screen for everything else.
- The top row has the buttons for the clip you're watching: Edit and Create as is. After you create the clip, they change to Download and Create post. When you're watching Full video, you see Add range.
- Tap Clip list to switch clips, or select several clips and create them at once. When you tap a clip in the list, the list closes and that clip appears.
- Tap Settings to change the settings above. Your changes stay when you close the settings.
Move a clip to the editor
For more detailed edits, click Edit to move the clip to the editor. To learn what you can do there, see Video editor and templates.
Once you move a clip to the editor, you can change its settings and create it only in the editor, and you can't go back to the settings in the list. Set the template, silence removal, and whether to follow the speaker before you move it.
Make it in another language
To post the same clip in several languages, use Copy & translate. It keeps the original audio and the range, and creates a new clip (a language version) with the captions, title, and comment cards translated into the language you pick. The original video isn't analyzed again, so the clip stays the same.
Choose languages and clips
Click Copy & translate in the button row below the preview. Language versions are made from the clip settings, so this button isn't in the editor. Choose one or more languages under Translate into, then under Clips to translate choose This clip or all original clips. You get one language version per clip and language.
Check the language versions
A language version appears right under its original clip in the clip list, with the language name. Click it to check the translated captions in the preview. You can fix the title and comment cards in the settings on the right.
Fix the captions
Click Edit to move the language version to the editor. In the Captions tab on the left, each sentence shows the original above it, so you can compare and fix the translation.
- Translated captions appear while each sentence is spoken, and caption effects such as word highlighting look the same as on the original clip.
- The original voice stays. To dub a language version, move it to the editor and use Replace original voice in the Voice tab. A voice for that language is already selected. To keep the captions short while the voice reads the full translation, write a separate voice script.
- Changes made in the editor (caption text, text, effects, music) aren't carried over. Language versions are made from the clip settings.
- If you set a Long-form → short-form template for that language in Localization, language versions use it instead of the original template. It's used only when its frame fits the clip's source video. The Localization box at the bottom of the window shows which template each language will use.
- You can't change the range of a language version. Change the range on the original clip, then create the language version again.
- If your brand rules list how a name is written in each language, the translation uses that spelling.
- For videos with several characters, add each person's gender and relationships to your brand rules. Captions are translated without knowing who is speaking, so a person whose gender the lines don't show can end up with the wrong word, such as he or she.
- Switch between the original and its language versions with the language menu next to the project name or next to the title in the editor.
Cancel an analysis
While an analysis runs, you can stop it with Cancel analysis on the progress screen. We refund the credits it used.
Credits
- The analysis costs a base amount plus credits for each minute of the original video, and you see the estimated cost before you start.
- Each clip you create uses credits. Clips longer than 90 seconds, removing original subtitles, and following the speaker cost more.
- Copy & translate uses credits for each language version it creates, and creating the video for a language version is charged separately.
- If an analysis fails or you cancel it, we refund the analysis credits. If you cancel a render while a clip is being created, we refund those credits too.
To see how many credits each feature uses, open the Credit usage guide on the Credit usage page.