Flaming Rider is a one-tap file, not a prompt you rewrite weekly
/templates/flaming_rider is a photo in, a clip out. A new generate is a different object.
flaming_rider is a photo in, a clip out. A new generate is a different object. The Flaming Rider template is not a prompt you rewrite every Monday when the sky feels stale. It is a named one-tap file: upload the still, run the format, get the clip. The vintage motorcycle, the city, the flaming-skull transformation — those choices already happened when the template was built. Your job is the photograph.
Rewriting a Veo prompt to "do Ghost Rider again" is a generate. It will invent a different rider and a different bike. That is not a rerun of the template. That is a new object you will then try to match by hand.
The template is a file, not a prompt library
Flaming Rider sits in the trending catalog. The description is the format: put yourself on a vintage motorcycle in the city, then burst into a flaming-skull Ghost Rider transformation. The only required input is user_image_url, labelled "Your photo." There is no scene list. There is no "describe the highway this week."
Run a one-tap AI video template is the agent job. get_template tells the agent which inputs that id needs. Once the photo is a hosted URL from the attach button, start_template_run starts it. A live card tracks the run. get_template_run and cancel_template_run support that card. You do not poll in the ordinary case. You do not pick Veo versus Kling inside the template. The template already chose its stack.
Estimated cost is on the template before you run it: about 89 credits, rising to about 129 for a full-quality render. That quote is the object. A blank video generate with a 400-word rider prompt is a different quote and a different failure mode (the face will drift).
The named-job writeup is a one-tap AI video template is a photo in and a finished clip out. This page is only flaming_rider.
Weekly rewrites are a generate you did not need
The instinct is understandable. Episode 12 wants rain. Episode 13 wants a desert. So you open a text-to-video box and start specifying skies. That is how you leave the template. The job is the format — bike, city, skull-and-flame transformation — applied to this still.
If you need a different format, pick a different template from the templates catalog. If you need a custom bike or a multi-scene episode, that is a movie or a generate. What you can change every week: the photo, the caption after export. What you cannot change by rewriting a prompt: the template's blocking. There is no field that turns Flaming Rider into a product demo.
Starting before the photo is a hosted URL is a failed run. Attaching a group shot and hoping the template meant the person in the back is how you spend the estimate on the wrong subject. Face-on, sharp, one person.
A new run is a new object
Each start_template_run is its own file. It does not continue last week's clip. If the new still is a different angle of the same face, you will get another rider take of that still — same format, new generate. Treating two runs as "one series you are iterating in the prompt" is how people open a custom generate and never come back.
Cancel an in-flight run with cancel_template_run. Closing the chat is not a cancel.
If no template matches, generate. Forcing Flaming Rider to be a helmet review or a product orbit is paying for the transformation format and then not posting it. The format DNA is a separate post. This one is only the object: photo in, clip out, stop rewriting.
FAQ
Do I pick the video model inside Flaming Rider?
No. The template already chose its stack. If you wanted to pick a model per shot, you wanted a generate or a movie, not one-tap.
Can I describe this week's sky in the prompt?
Not on this object. The template takes your photo. Sky direction is a different job. Run the template, or leave it and generate. Do not mix the two in one estimate.
What photo does it actually need?
Your photo, required. Clear, front-facing, one person. That is the still the format consumes. A group shot, a tiny crop, or a photo of the bike instead of the rider is how the run misses.
Is a new weekly run "the same video, updated"?
No. Each run is a new file from the current still. Same format, new object. That is the point of one-tap. It is also why you should not spend a week rewriting a custom prompt to chase last week's flame.