How to prepare a Smalltown Boy dance video from two photos
A practical two-photo workflow: assign A and B, avoid role confusion, review the whole dance, and save a vertical result.
Updated
Prepare your two photos →Choose and swap two portraits in the creator, with Photo A opening the reaction and Photo B beginning the dance. The full video flow covers progress, 9:16 playback and MP4 download. Public generation is temporarily unavailable.
Start with two people who agree to take part
Use one clear photo per consenting adult. A group picture leaves the intended subject ambiguous; two photos give each performer a separate identity reference. Do not use a celebrity, a stranger’s profile picture or a friend’s image without permission. Choose the original files, review both at a useful size, and make sure neither face is covered. Our local chooser accepts JPG, PNG and WebP up to 10 MB each. Selecting a photo is a browser preview, not an upload. Closing or refreshing the page clears the selected photo files.
Assign the opening roles, then check the order
Photo A is the intended first foreground reaction. Photo B is the intended dancer behind. Both later take part in the exchange and shared dance; A and B describe identities, not permanent left and right positions. Use Swap A / B to reverse the cast without choosing the files again. Inspect the thumbnails before continuing. Clothing, posture and the timing of the final sequence may vary, so the assignment is a guide rather than a guarantee of exact reproduction.
Understand the task before starting another
When generation becomes available, start it deliberately and wait for the task status. The current restricted test interface distinguishes photo preparation, submission, checking a saved task, processing, an uncertain provider response, failure and a playable result. It does not pretend to know a completion time. An interrupted connection does not mean the provider stopped. Recheck the existing task rather than submitting a duplicate. A saved task reference can be restored in the same browser tab and account when browser storage is available; this is not a permanent video library. Selected photo files are not restored after refresh.
Watch the entire vertical frame
Review the video from beginning to end, including both reactions, the role change and the wider dance. Check that each face belongs to the intended person and that limbs and hands remain plausible. Do not judge quality from a poster frame alone. The result player keeps the full frame visible instead of cropping it to fill the panel. Watch larger opens a focused player with native playback controls. Close it to return to your cast and task; playback position is retained. No real generation is needed to prepare your photos today.
Save, review and share responsibly
A completed result provides an MP4 download. On a phone, the file may go to the browser’s Downloads or Files before you save it to Photos or share it. Keep a local copy before deliberately creating another result; public history is not currently available. Review likeness and permissions before sharing, and disclose AI editing where appropriate or required by the platform. Do not assume a downloaded file includes licensed music: choose audio you are permitted to use for the intended platform, territory and account. Announced one-time packs are $9.99 for 1 video, $19.99 for 3 and $29.99 for 5. Checkout is not active yet. The published output format is 20 seconds, 480p and vertical 9:16; no free public trial is offered.