Put an original character into the Bulin 47 performance. Choose your clip length and frame, then use the prepared reference to guide the scene.
Original reference · not an AI resultBulin 47 performance studio
Character photo is all you need.
One clear face · JPG, PNG or WebP · max 10 MB
Your character reference and the original video guide the performance. The matching source audio is restored afterward. Appearance, gestures and lip sync may vary.
Template preview · generation and checkout are paused.
Performance preview
Kevin Wai’s original recording · 8s reference, not an AI result. View source ↗
Completed takes will be shown here when rendering opens.
One photo in
Original referenceThe character image should clearly show a face, hairstyle or fur pattern, and the intended outfit. It supplies the appearance of one lead rather than an identity for every person in the scene.
Original referenceThe selected excerpt provides gesture timing, camera movement and foreground occlusions. Instructions target the performer behind the speaker holder, while asking the model to retain the other roles.
Original referenceThe 8-second option is a compact take; 15 seconds includes more of the sequence. The 23-second option joins two sections before restoring the source soundtrack. Check likeness, hands and sound before sharing.
Images show the original reference. AI may change background details and surrounding people. Read the template guide ↗
Three simple steps
01 / 03
Choose an original character image or a photo of an adult who authorized this use. Keep the face and clothing easy to read.
02 / 03
Select 8, 15 or 23 seconds and the output frame. The reference clip supplies the performance rather than a new prompt-only scene.
03 / 03
Review the character and motion result before downloading. Rendering is paused today; you can preview the source and keep a local setup.
Pricing
2 credits
4 credits
10 credits
20 credits
Good to know
The intended replacement is the lead behind the portable speaker. Your image guides that character’s appearance; source footage guides staging and motion. AI can still alter nearby people, fingers, clothing and mouth timing.
No new imitation of your voice is requested. The delivery stage restores the selected source recording, which still requires the relevant audio and performance permissions.
The current release is a preparation studio. It plays the source excerpt, previews your selected image and saves settings locally. Public model submission and payment are disabled.
All options begin at 00:09 of Kevin Wai’s recording. Eight seconds ends at 00:17, fifteen ends at 00:24, and twenty-three ends at 00:32. The longest option is processed as 15 seconds plus an 8-second continuation.
Use a single, well-lit subject with enough face and upper-body detail to identify the intended look. Avoid heavy filters, a covered face and group shots. Only submit material you own or are allowed to process.