From Still Image to Video: When to Use Which

The default should be the cheaper one

Video generation costs more than image generation in every dimension that matters: credits, time to render, time to review, and time to fix when it comes back wrong. A 15-second clip costs 30 credits; a still costs one, and the first three each month cost nothing at all. That ratio should influence what you reach for first, and for most creators it doesn't.

The useful question isn't "would this be better as a video?" — almost anything would. It's "does the motion carry information the still doesn't?" If the answer is no, you're paying thirty times more for decoration.

What motion is actually for

Motion earns its cost when the thing you're communicating is a change: a process unfolding, a before and after, a character reacting, a sequence with an order that matters. In those cases a still can only summarise what a clip demonstrates, and the summary is genuinely worse.

Motion does not earn its cost when the content is a state rather than a change. A character introduction, a product shown at rest, a title card, a mood-setting establishing shot, a thumbnail — these are states. Animating them produces a clip where nothing happens slowly, which reads as padding even when it looks good.

The cases where a still is clearly right

Thumbnails and covers. These are consumed as stills by definition. Generate them as stills, at the ratio you publish in.

Character design and cast building. You're settling what someone looks like, not what they do. Do this in images, where the iteration loop is seconds instead of minutes.

Storyboards. The entire point is to be cheap enough to throw away. Boarding in stills before rendering is the single most effective way to reduce wasted video credits, because it surfaces the scenes you don't need while they still cost one credit each.

Social posts that aren't video-first. A still that reads instantly often outperforms a four-second clip that has to be watched.

The cases where video is clearly right

Anything with a sequence. If the beats have an order and the order matters, that's video.

Explanation of a process. Showing the steps in sequence is the whole value; a still of step three explains nothing.

Storytelling formats. Story channels, faceless narration, and nursery-rhyme content are built on duration. A still is not a smaller version of these — it's a different product.

The hybrid that works better than either

The most economical workflow uses both, in order. Board the piece as stills. Cut the beats the board reveals to be redundant. Then generate video only for the beats that survived — and use the approved still as the opening frame of its scene, which is the most direct control you have over how a clip begins.

This matters more for longer pieces than short ones. Because videos over 15 seconds are produced as chained scenes, where each clip continues from where the last one ended, the opening frame of each scene is where continuity is won or lost. A board you already approved is a far better starting point than whatever the model would have invented.

A rough decision rule

Ask what the viewer is supposed to do with the thing. If they're meant to glance and decide — click, scroll, recognise — it's a still. If they're meant to follow along — watch, learn, be told a story — it's video.

Most channels need more of both than they think, but in a specific ratio: a handful of videos, and many more stills supporting them as thumbnails, boards, and character references. Budgeting as though everything must be a video is the most common way to run out of credits with nothing published.

Where the free quota fits

Three free images a month is enough to board a short video, or to settle a character, or to test two thumbnail directions — but not all three. Spending it on the most consequential question and then paying for the rest tends to beat spreading it thin. The pricing page has the per-credit detail for both halves.