Desert Stunt is a one-tap file, not a prompt you rewrite weekly

desert_stunt is a photo in, a clip out. A new generate is a different object.

The one-tap template sits at Desert Stunt. The id in the product is desert_stunt. One required input: Your photo (user_image_url). The format is already chosen: sit unbothered on a desert chair while a black off-road truck launches off a dune and flies clean over your head. Estimated 68 credits, about 135 for a full-quality render. You do not rewrite that sentence every Monday. You attach a still and run the file.

A prompt you baby-sit in the AI video generator is a different object. New prompt, new model row, new miss. Templates are not that loop.

The template already chose the shot

One-tap means the creative choice happened when the template was built. Camera, gag, truck, chair, grade — those are not knobs you re-derive from a weekly brief. Run a one-tap AI video template is the named agent job: start_template_run after every required input is a hosted URL. get_template lists what that template needs. A live card tracks the run. You do not poll, and you do not paste a new screenplay into the template name.

If you wanted to pick Veo versus Kling, write a three-act, or change the vehicle, you wanted a generate or a movie. The template will not take a scene list. It will take the photo.

This page is not a remake of the trend explainer. The desert stunt trend post already covers spectacle and casting. This page is only the object: a saved format, not a prompt.

Your photo is the only input

desert_stunt has one field, and it is required. No optional second still. No default character if you skip the upload. Attach via the chat + button so the input is a hosted URL, then start. Launching with an empty inputs object is a failed run, not a clever prompt.

The still is the only place your taste belongs. Face visible, one subject, light enough to survive a bright desert grade. A group shot is how identity blends. A weekly rewrite of "make the truck redder" is not an input the template exposes. If you need a different gag, you need a different template, or you need to leave one-tap.

Credits are estimated on the template page before you run it. Cloning a public workflow is a different agent job and a different cost shape. One-tap is photo → this clip.

A new generate is a different object

People burn the template by treating it as a seed for a custom film:

• Rewriting the prompt in the generator every week. That is text-to-video or image-to-video. You left start_template_run. You are now paying a model row to re-solve a shot the template already solved.
• Asking the agent to "do desert stunt but night / two trucks / a line of dialogue." The template does not have those parameters. You will either get the stock gag anyway or a generate that is not the template.
• Cloning a workflow and calling it one-tap. Saved workflows rewrite plots. desert_stunt does not write plots.

Run the file when you want this clip with a new face. When you want a new clip, pick a new object. The templates catalog is the shelf. The generator is the blank page. Mixing them is how a 68-credit format turns into an open-ended spend you will not reproduce next week.

FAQ

Do I pick the video model inside Desert Stunt?

No. The template already chose its stack. You supply the photo. Model shopping is a generate.

Can I change the truck or the chair in the prompt?

Not inside one-tap. Those are the template. A new generate can try; it is not desert_stunt, and it will not cost the template's estimate.

What does it cost?

The catalog estimates 68 credits, about 135 for full quality. The agent shows the template's estimate before you run. Do not convert that into a dollar price.

Is this the same as recreating a public workflow?

No. Recreate clones a multi-scene workflow into your account, saved, not auto-run. Run a one-tap template starts desert_stunt (or another one-tap id) against your photo and returns that template's finished video.