Pretend To Sleep is a one-tap file, not a prompt you rewrite weekly

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

The template already chose the gag: your pet gets caught scrolling in bed and frantically pretends to be asleep when the owner opens the door. You do not re-prompt that timing every Monday. You attach a pet photo. You run the template. Rewriting a Veo paragraph each week is a different object — a generate — and it will miss the hold-and-twitch the template encodes.

One-tap is a template id

The in-app id is pretend_to_sleep. The page is Pretend To Sleep. Required input: one image, labelled "Your pet photo." Category: Trending. Estimated cost: about 68 credits, rising to about 135 for a full-quality render. That estimate is listed before you run it.

Run a one-tap AI video template is the named agent job. get_template tells the agent which inputs that id needs. Once every required input is a hosted URL (attach via the chat + button), start_template_run kicks it. A live card tracks progress. get_template_run and cancel_template_run support that card. You do not pick Veo versus Kling inside the template. The template already chose its stack.

Output is that template's finished video, built from your upload. No timeline. No scene list. No "write me a three-act structure."

Rewriting the prompt is a different object

A weekly generate — new paragraph, new model, new timing — is generate_video wearing a trend costume. It can land. It often does not, because Pretend To Sleep is a timing joke: still, twitch, snap back. The template encodes those beats. Your prompt will argue with them.

A new generate is also a new bill shape. Templates list an estimate per run. A blank generate bills the model you picked, at that model's duration and resolution. Using a generate to recreate a one-tap format is how you pay video-model rates to rediscover a gag that already has an id.

Do not clone a public workflow and call it Pretend To Sleep. Recreating a viral workflow (recreate_public_template) saves a scene graph into your account. It does not auto-run. One-tap start_template_run is photo in, clip out. Different tool, different object.

Do not start the run before the pet photo is a hosted URL. A missing still is a failed run you still have to notice.

The still is the only weekly work

The cadence that matches the product:

1. New pet photo, or a new angle of the same pet — clear, front-facing.
2. Open Pretend To Sleep (or ask the agent to run that template).
3. Attach. Run. Watch the card.
4. Post the file. Next week, a new still. Same id.

That is a series. It is not a prompt library. If no template matches the gag you actually want, generate. Forcing Pretend To Sleep to "just be a product ad" is how you pay the template estimate for a joke format you will not post.

If you needed a multi-scene story, make a movie. If you needed to pick the model per shot, generate. Pretend To Sleep is one named template. Using it as a weekly prompt scratchpad wastes the only thing the template is for: the file, on time, from a photo.

FAQ

Can I rewrite the template's story each week in the prompt box?

Not as this object. The creative choice already happened when the template was built. You supply the pet photo. If you want a new plot, that is a generate or a saved workflow, not start_template_run on pretend_to_sleep.

How much does Pretend To Sleep cost?

The catalog lists an estimated 68 credits, about 135 for a full-quality render. Each template shows its own estimate before you run it. That is not a blank Veo 8-second bill, and it is not zero.

What photo does it need?

One required image: your pet photo. Upload a clear, front-facing still. The template will not invent a pet from a paragraph.

Is running the template the same as generating a new video from a prompt?

No. A generate picks a model and a prompt. A one-tap template applies a format to your photo. A new generate is a different object even when the output is also an MP4.