Browse all docs

Turn a long video into YouTube Shorts automatically

PandaStudio's Make clips turns one long video into scored 9:16 Shorts with captions and face-following framing, finding the moments with an AI model that runs on your own computer.

Drop a podcast, webinar, lesson or stream into Clips. PandaStudio transcribes it, reads the whole transcript to find self-contained moments that open on a hook and end on a finished point, and turns each into its own vertical project: captions in the style you picked, the frame following the speaker's face (or the whole screen over a blur when there's no face). Clips come back best first with a title, the opening line and a 0 to 100 score. Want clips about one subject only? Type it in What should the clips be about? before you drop the video (for example voter registration, or the pricing part) and only the moments about it become clips (PandaStudio 2.1.52 and later). If nothing in the video is about it, you get the strongest moments overall and a note saying so. Every clip is a normal project, so you can trim it, restyle it or hand it to the agent, then export them all at 1080x1920. Nothing is uploaded and there are no credits or minute limits. In PandaStudio 2.1.61 and later, each clip card has a Publish button that exports the clip if needed and opens YouTube and social publishing, where a vertical clip of 3 minutes or less goes up as a YouTube Short, and the Clips sidebar shows each video's length, even while it is still importing or processing.

A six-minute video turned into captioned Shorts, plus a two-person podcast stacked, in the real app.

In the app

  1. 1On Home, click Make clips in the What do you want to make today? box (or open Clips in the sidebar) and click New long video.
  2. 2Optional: in What should the clips be about?, type a subject in a few words, such as voter registration or the pricing part. Only moments about it become clips. Leave it empty for the strongest moments on any subject (PandaStudio 2.1.52 and later).
  3. 3Pick a clip length (15-30, 30-60 or 60-90 seconds) and a format (9:16, 1:1 or 16:9).
  4. 4Drop in your long video, or click to choose it. Processing starts straight away.
  5. 5While it works, pick a caption style: each tile shows the style on a frame of your own video. You can change it later and every clip is restyled.
  6. 6The clips come back best first, each with a title and a 0 to 100 hook score. Hover over one to play it. A podcast shot with two people in frame is stacked, one person above the other.
  7. 7Click Edit on any clip to polish it in the editor, or Export (Export all for every clip).
  8. 8Click Publish on a clip's card to publish it to YouTube Shorts or post it to Instagram, TikTok and more. If it isn't exported yet, it exports first (PandaStudio 2.1.61 and later). In the Exported tab, Play & publish does the same for clips you already exported.
Home screen with Make clips outlined in the start box
On Home, click Make clips in the start box.

Ask the AI

Open the AI chat in the editor and say any of these. The agent does the edits for you.

“Make Shorts from ~/Videos/podcast-ep12.mp4 with short 30-second clips.”
“Turn this webinar recording into clips with Hormozi-style captions and export the top three.”
“Find the best moments in this talk, then polish the top clip in the storytime style.”
“Make clips from ~/Videos/town-hall.mp4, only the parts about voter registration.”

CLI & MCP

Every capability is a verb, reachable from the terminal or from an MCP agent (Claude Code, Cursor, Cline, Continue, Claude Desktop).

clips.make
pandastudio clips.make --file=/path/to/podcast.mp4 --length=medium --captionTemplate=glowStack

Returns { clipSetId, jobId }. length: short | medium | long. footage: camera (default) | screen. captionTemplate: any caption template id, or none. topic (2.1.52+): only moments about it, e.g. --topic="voter registration"; clips.get shows topicMatched: false when nothing was.

clips.get
pandastudio clips.get --id=$CLIP_SET_ID

Status and the clips best first: title, hook, score, startMs/endMs and each clip's projectPath.

clips.export
pandastudio clips.export --id=$CLIP_SET_ID

Exports every finished clip (or --clipIds) to Movies/PandaStudio Clips/<video>/.