What Kinds of Videos Can You Make With Visuall?

Updated

Visuall is built for two ends of the same spectrum: long-form campaign films and short vertical clips, both made on one canvas. According to the site, you can make a product film, a vertical Reel, a short film, or a feed-ready eight-second clip — and you approve every keyframe before it gets animated. If your workflow depends on reviewing stills before committing to motion, or on changing one shot without re-rendering the rest, this is the fit. If you need live-action footage or a fully automated one-click output, it is not.

The formats Visuall is designed around

The homepage frames the range directly: "AI video, from campaign film to viral clip. One canvas for the campaign film and the eight-second clip." The examples shown on the page cover:

  • Product film — a horizontal, campaign-style piece
  • Reel — vertical, "made for the feed"
  • Reel — first-person action, e.g. "Fresh powder, first person," shot vertical
  • Product film — built from three phone photos
  • Short film — a narrative piece, e.g. "A day in Rome, 44 BC"

So both aspect ratios are in scope: vertical for social, horizontal for campaign and narrative work. The page does not list a fixed maximum length, so treat duration as something you set per project rather than a published limit.

The workflow that shapes what you can make

The canvas model is what determines the kinds of videos that are practical here. Every video is "a set of steps you can see and change": your story, the AI scene plan, and the model chosen for each shot. The key constraint-turned-advantage is stated plainly — "Approve the frames you like before they are animated, and re-run only what changed; the rest is reused for free."

That means the format best suited to Visuall is anything where you want shot-level control:

  • Multi-scene narrative — a story broken into scenes, each with its own keyframe and animation
  • Brand or product work — where a wrong frame is expensive and you want sign-off before motion
  • Iterative social content — where you tweak one shot and keep the other seven

The page's own example walks through a 10-second, 3-scene skier sequence: an aerial dolly over a snowy ridge, a low tracking shot with powder spraying past the lens, and a wide shot of the skier carving into the valley. Each scene gets a keyframe (FLUX.2, 16:9), an approval step, then animation (Kling 3, 5s). That structure — story → scenes → keyframes → animate → stitch — is the template for anything you build.

A concrete example: the 44 BC Rome short film

The site's behind-the-scenes breakdown is the clearest illustration of what a finished Visuall video looks like. It describes "a vlog from Rome in 44 BC, by someone who doesn't exist," made in four steps across 11 boxes:

  1. Describe her once — one paragraph about a character named Mia; FLUX.3 draws a consistent face used in every shot
  2. One line per scene — four places in Rome, one per line; Edit Image places the same face in each
  3. Write what she says — Gemini Omni Flash turns each frame into an 8-second clip that talks, with sound
  4. Stitch the episode — Stitch joins the clips into one 32-second video with her voice included

The page notes this took 225 credits and "about two minutes of waiting." That gives you a realistic sense of scale: a 32-second narrative short with dialogue and sound, built from four scenes, is a normal-sized Visuall project.

What this means for your choice

Pick Visuall when your video needs scene-level approval and selective re-rendering — campaign films, product pieces, narrative shorts, and vertical social clips where you want to fix one shot rather than regenerate everything. The canvas, the per-shot model selection (Veo 3.1, Seedance 2.0, Gemini Omni, Kling 3, FLUX.2/FLUX.3 all appear on the page), and the "only what you change is made again" behavior are the reasons to choose it.

Look elsewhere if you need real captured footage, or if you want a single prompt to produce a finished video with no intermediate approval step — the whole design here assumes you want to inspect and approve frames along the way.