Gas Station AI Trend: how the three-photo workflow works
A practical guide to the current Gas recipe, its limits and the waitlist.
Start with the availability notice
Framecomet's public Gas generator is on a waitlist. An illustrated scene on the page is not proof of a completed generation. Check the availability notice before preparing a purchase; public checkout is also closed while acceptance is in progress.
Assign one adult to each role
The current recipe expects three separate portraits. Assign the first person as the dancer, the second as the selfie walker and the third as the person beside the car. Use different consenting adults and keep a note of which image belongs to which role. A single group photo does not replace these three inputs.
Understand the current recipe
The internal recipe creates a first frame with GPT Image 2 and then animates it with Seedance 2.0 mini. The validation settings are 10 seconds, 720p and a vertical 9:16 frame. These describe what is being tested; they do not guarantee exact identity, precise choreography or three separate camera shots.
Review before you share
When the service opens, review the credit cost before confirming. A result must finish processing and safety review before it can be released for download. Inspect faces, hands, the three roles and unwanted changes yourself. There is no promised soundtrack: add music only if you have the right to use it.
Read the photo preparation guide and failure and credit guide, or return to the Gas page to join the waitlist.