Edit Narration and Visuals

Magic Video, Shorts, and Voiceovers use the same Video editor. It keeps narration, generated audio, visuals, and timing in sync, and lets you change each part on its own.

Edit Narration

The narration editor contains the complete spoken script, divided into paragraphs that match the generated audio segments. Text changes save automatically, but they do not silently replace finished audio.

After changing a paragraph, use Regenerate Audio on that paragraph or regenerate every stale paragraph from the narration toolbar. Every paragraph shows its Host. In Magic Video and Voiceovers, choose another Host to change that paragraph, or leave Swap All Hosts? on to exchange both Hosts everywhere they appear in the script. A Short uses one narrator, so choosing a Host there moves the whole script to that Voice. Host changes mark only the affected paragraphs as stale until you regenerate them in the new Voice.

You cannot render, download, or update the shared version while narration is stale. This prevents a finished Video from combining new text with old audio.

Edit Visuals

The Visuals lane, the first lane in the timeline, shows each shot in order as a strip of thumbnails. Click a visual to select it and seek to its moment in the Video.

Right-click a visual, or select it and choose Edit Visual, to open a single dialog where you can:

• Regenerate: Describe what should change, or leave the direction blank for a fresh interpretation. The dialog shows the cost, 7 credits per regenerated visual, before generation starts.
• Replace: Drop your own image or video onto the same dialog, or drag supported media directly onto a visual thumbnail in the lane.

If the image model declines a shot for its content, regenerating won't change the outcome, so Jellypod says so directly instead of retrying: edit the description, then regenerate.

Voiceovers also get a separate Reframe option: open the supplied page, image, or extracted figure in a larger editor, drag it beneath the fixed Video frame, adjust the zoom, show the full visual, or reset to the automatic framing. Saving changes only this appearance of the visual.

Right-click a visual and choose Delete to remove it; a confirmation dialog appears first. The neighboring visual extends to fill the gap, and the change can be undone.

Regenerating, replacing, or deleting one visual leaves its narration and timing unchanged. Rendering is unavailable while a visual is regenerating.

Timeline

Narration and music sit in their own lanes beneath the Visuals lane, on the same track-based Timeline Editor used by AI Podcasts: drag, trim, cut, copy, duplicate, and delete audio segments, adjust volume and fades, and use zoom to see more detail or the whole Video at a glance. Jellypod automatically reflows visual timing when narration audio changes, so visuals stay lined up with the words they illustrate.

Product Differences

• Magic Video and Voiceovers: May use one to four Hosts, assigned per paragraph in the editor. Magic Video's generated script assigns the Hosts for you; a Voiceover is narrated by its first Host until you reassign paragraphs.
• Shorts: Use one Host for the complete script.
• Magic Video and Shorts: Use AI-generated visuals in the Visual Style selected before creation. The style and orientation cannot be changed after generation begins.
• Voiceovers: Begin with creator-supplied visuals and always use landscape orientation. The creation flow has no Visual Style, Brand Kit, watermark, or background music choice; the shared editor lets you add background music or a watermark afterward.
• Shorts: May include an instrumental score that you can replace or remove.
• Magic Video and Shorts: May include a Brand Kit, watermark, and captions during creation. Voiceovers support captions and an editor-added watermark, but not a Brand Kit.