Image to video: turning product photos into scroll-stopping ads

Image mode turns a stack of stills into a moving, narrative video. The ordering matters less than you think — Agent 2 (the rhythm director) will reorder, reuse, and even split your uploads to maximize narrative flow. This article covers what to feed it and what to expect.

What Agent 2 actually does with your images

Once you upload images and hit Start, Agent 1 looks at every frame and writes a short visual summary. Agent 2 then reads those summaries plus your prompt and produces an editing plan: which image becomes the hook, what motion each one gets, how long each beat runs, and where the trims sit.

• Reordering is fair game. The strongest visual is often a better opener than the first upload.
• Reuse is allowed. The same image can come back as a callback — Agent 2 will give each occurrence a different motion.
• Wide images can be split. A 21:9 landscape can become two distinct shots (left half + right half) with different motion.

What makes a good upload set

Variety beats volume. Five strong, varied images produce a much better cut than thirty near-duplicate selfies. Aim for a mix of:

• One hero shot — the most visually striking image
• Wide / establishing shots that set scene
• Detail shots (textures, hands, faces)
• A “payoff” — the visual that pays off the hook

The Hook card matters most

Image mode honors the Hook card carefully. Set it to Bold and Agent 2 will lead with your strongest single image. Set it to Build-up and the strongest image is held back for scene 3 or 4, while the opener stays calm. Reveal may even crop a portion of the strongest image as a teaser and show the full thing near the end.

Iterating with Tweaked

Your first run is the baseline. Change any Visual Adjustment cards and click Start again — Vidmonto generates a Tweaked version next to the baseline so you can compare. Tweaked runs are billed at the same rate as the baseline.