Turn one full video into clip candidates you can review.

How it works

Auto Clips is the Pro workflow for starting with one long recording instead of a folder of finished edits. You choose the maximum clip length and how many clips you want. ClipPush then selects the moments, follows the active speaker and adds the finished 9:16 clips to your library.

1. 01

   Open Auto Clips

   Choose Auto Clips in the dashboard. Create an episode first if the source does not already have one, then select where the finished clips should be kept.
2. 02

   Upload the full video

   Choose your MP4, MOV or WebM in Auto Clips, not in the Clips batch uploader. Keep the page open until the upload itself reaches 100 percent.
3. 03

   Choose the Auto Clips settings

   When the transcript is ready, choose a maximum length of 30 seconds, 1 minute or 2 minutes, then choose between 1 and 10 clips. ClipPush reduces the available count when the source is too short for the combination you selected. You can also let ClipPush frame speakers automatically or place each speaker's frame yourself.
4. 04

   Set manual speaker frames

   Choose Manual if you want full control. ClipPush shows a representative frame for each voice in the transcript. Type the person's name, drag the yellow vertical frame over them, then add another speaker when needed. The finished clips will cut between those fixed positions as the speaker changes.
5. 05

   Make the clips

   Select Make clips and let ClipPush find complete moments, identify who is speaking and render the vertical versions. You can leave the page after the upload has finished.
6. 06

   Review the results

   The finished Auto Clips appear in your clip library and stay linked to the original episode. Open them to edit, choose thumbnails and add the strongest clips to a campaign.

What to know

• Auto Clips is available on Pro and Teams.
• Source videos can be up to two hours or 5 GB. MP4, MOV and WebM are supported.
• Landscape 16:9 footage works best. Auto Clips creates 9:16 output and uses clean cuts when a meaningful speaker change occurs.
• Manual framing uses the timecoded transcript to cut between the speaker positions you set. If a voice is not marked, the previous safe frame is held rather than guessing a new position.
• The selected length is a maximum. A clip can finish earlier when that creates a more complete thought.
• The original full video is private and is removed after seven days. The generated clips remain in your library.
• The Auto Clips page shows upload, transcription, selection, reframing and completion progress. A failed source keeps its error message so you can see what needs attention.