The free preview answers one question

Most free tiers are the paid thing with a ceiling on it: fewer renders, lower resolution, a logo in the corner. This one is not shaped like that, and the difference matters before you spend it rather than after. The preview is a single question put to a single product, and the parts of the product that would answer other questions are switched off at the server rather than hidden in the interface.

That is worth knowing in advance for a plain reason. If you believe you are holding a small version of the finished thing, you will read conclusions off it that it cannot support — about the length, about the crops, about what you could fix afterwards — and some of those conclusions will be wrong.

What it is

One render of a photograph you upload, free, with a watermark burned over it, and one of them per account. The watermark is not a nag; it is the thing that keeps the render a preview rather than a free film, and unlocking the clean master of that same render costs what generating it would have cost, with no second render and no second wait.

The important half of that sentence is "a photograph you upload". It is not a demo reel of somebody else's jewellery. The question it exists to answer is whether a camera moving across *your* object, lit the way Nitaris lights it, produces something you would put on a listing — and that question cannot be answered by looking at a gallery, because the gallery is full of objects chosen because they film well.

The seven things it will not do

Each of these is enforced on the server, not merely absent from the screen, which is why none of them has a workaround:

• It is always one ten-second take. The Studio still shows the length picker when you have a preview available, with the hint changed to say previews are always ten seconds — and if a request arrives asking for forty anyway, the length is read and thrown away. A forty-second film is four provider renders, and giving four away is a different business.
• It comes in the master ratio only. The square and 4:5 placement cuts a paid film carries are not produced while the render is locked, so the preview cannot show you how the listing crop will look.
• It cannot be trimmed or re-cropped. Both are free on a paid film and both refuse on a preview with "unlock the film before editing it".
• It cannot be extended. Adding ten more seconds to the end is a paid operation on a paid film and is not available here at all.
• Its last beat cannot be retaken. Same answer, same reason.
• It cannot be published to a connected account. Publishing is for files you own outright, and a watermarked preview is not one.
• It cannot be disputed. The fidelity check exists to give credits back when a render does not hold the product true, and no credits were charged for this one, so there is nothing to reverse.

There is an eighth that is not a limitation so much as a boundary: the preview is a browser thing. A shoot started inside the Shopify admin is refused outright, because a merchant's economics there run through Shopify's billing rather than ours. If you install the app first and have never used the browser, you have not spent your preview — you simply have not been offered it yet.

Which leaves exactly one question

Strip those away and what you are left with is the only question worth paying to answer early: does this object survive being filmed? Not whether the film is the right length, not whether the crop works on a listing, not whether you can fix the ending — all of that is decidable later, and some of it is free later. Whether the surface of your product does something interesting when light moves across it is not decidable later. It is the one fact that determines whether any of the rest is worth buying.

So the preview is pointed at that and nothing else, and the honest way to read it is: if this looks good, the mechanism works on my product and everything after this is a choice about how much and what shape. If it looks flat, no amount of length, crop or re-cut rescues it, and you have found that out for nothing.

How to spend the one you have

Three decisions, and only one of them is reversible:

• Pick the hardest product, not the best-selling one. You are testing the mechanism, and the mechanism is most informative where it is most likely to struggle — the glossiest, the most faceted, the most reflective thing you sell. If that works, the matte things will.
• Pick the photo you would actually start from. A preview generated from your cleanest studio shot tells you about that shot. If the pictures you really have are phone photos on a table, use one of those, because that is the input the rest of your catalogue will come with.
• Pick the ratio you will publish in. This is the one that matters most, because it is the only choice in the list the preview carries and cannot be changed afterwards: no placement cuts are produced while the render is locked, so the shape you choose is the only shape you will see. Choose the one you would put in front of a buyer, not the one that fits the screen you are sitting at.

And then watch it twice, once with the sound on and once paused on the first frame, because those are the two states it will actually be seen in.

What you still will not know

Worth being explicit, because this is where a reader could get it wrong. After a good preview you know the product films well, the mood suits it and the fidelity holds. You do not yet know whether a longer arc helps this object, whether the square crop keeps the label legible, or how the film reads after a trim — and none of those are things the preview was withholding to make you pay. They are questions a ten-second master-ratio locked render structurally cannot be asked.

The sequence that follows from that is unglamorous and cheap: preview the hardest product, unlock the one that works, and only then start spending on length and placements, with a film you already know is good underneath the decision.