Fan freestyle
A shoulder-led street performance.
15s · 720p: 15 credits · 1080p: 30 credits
Generate this videoLoading your scene…
Loading your scene…
Upload one photo. Choose an effect and video quality — we handle the motion and generation.
Checking your account…
JPG, PNG or WebP · up to 10 MB · each side over 340px
Follow progress and download your MP4 in My videos.
My videosGet credits (new tab)Upload privacy
Create an original street-freestyle scene with the consenting adult from the reference. Preserve their recognizable face and outfit. An adult partner holds a portable speaker at waist height. The performer looks toward the speaker, makes two alternating shoulder bounces, gestures to a loose spectator circle, then pauses. Medium-wide steady framing, visible hands, restrained background movement. No real-artist face or voice imitation, text or logos. A new interpretation, not original footage.
Your photo stays on this page until you generate. No free generation. AI results may vary.
Read the workflow guideWatch a template, then open its generator to create your version with one photo.
A shoulder-led street performance.
15s · 720p: 15 credits · 1080p: 30 credits
Generate this videoA longer street performance with hand gestures.
15s · 720p: 15 credits · 1080p: 30 credits
Generate this videoA short character-led motion. Silent reference.
6s · 720p: 6 credits · 1080p: 12 credits
Generate this videoThird-party examples, not generated by Bulin47AI. Each preview matches the selected motion template; generated results may vary.
The idea behind the trend
The name refers to AI remixes inspired by Bulin 47’s street freestyle with Kevin Wai. The recognizable ingredients are the shoulder-led performance, a nearby partner with a portable speaker, and a watching crowd.
A remix can put a new character at the center. A photo gives the character a look; a motion reference guides the movement. Those are separate inputs, and a text prompt alone does not reproduce the original performance exactly.
Keep the exchange.
Make the performer your own.
From photo to performance
Choose one clear character photo. Keep the face, shoulders and outfit visible. Use a photo you have permission to process; prepare an upright character image first for a pet remix.
Choose Fan freestyle or Performer remix for 15 seconds, or Character bounce for 6 seconds. We supply the motion reference; you only upload your photo.
Sign in with a verified email account. Choose 720p or 1080p, review the duration and credits, then generate. Follow progress and download your private MP4 in My videos.
One-time credit packs, no subscription. Choose 720p or 1080p motion transfer. No free generation allowance.
A little movie magic
Room to try a few ideas
For your whole cast
Choose 3–15 seconds per video. We reserve the quoted amount and settle against delivered duration, rounded up to the next second (50 ms tolerance). Shorter clips return the difference. You never pay more than your quote. Taxes, if applicable, are shown at checkout.
Watch. Plan. Recreate.
Upload one photo, choose a 6- or 15-second effect and select 720p or 1080p. We supply the motion; you see the credit cost before generating.
Read the guideThe most useful portrait clearly shows the face and the body area involved in your chosen motion. Match it to the reference rather than using a tiny face crop.
Read the guidePair your character photo with an authorized motion clip. Bring your own moves accepts MP4 or MOV, 3–15 seconds, up to 20 MB, with a 9:16 or 16:9 frame.
Read the guideA pet preset prepares an experimental idea. Human motion-control documentation does not establish reliable animal anatomy or successful pet output.
Read the guideAn original character works best as a coherent visual design with a movement it can plausibly perform. Keep the silhouette and visible joints readable.
Read the guideA useful prompt describes appearance and scene while the reference supplies movement. Keep the first brief short enough to review and revise.
Read the guideFind the stage where the problem happens before changing the prompt. A local preview issue, a provider failure, and an imperfect result need different checks.
Read the guideA usable export needs a checked video file and a checked audio track. Neither motion transfer nor a public video link establishes music or footage permission.
Read the guideBefore you start
Upload one photo and choose a fixed effect. We supply the motion reference. Choose 720p or 1080p, review the duration and credit quote, then generate with a verified email and paid credits. Bring your own moves remains available under Advanced workflows.
Yes. The default generator needs only your photo. Fan freestyle and Performer remix use 15-second templates; Character bounce uses a 6-second template. The platform supplies the motion reference and generation settings.
You can prepare those ideas here. A pet usually needs an upright stylized character image before human motion can be applied. Pet and mascot results have not been tested by this site.
The performer and speaker partner workflow has separate photo A and photo B slots for composition preparation. It does not send two identities into a single-character motion model or generate a duet.
Audio is not guaranteed. Character bounce uses a silent reference. Review the sound in the finished video and use audio you have permission to publish.
Preparing a scene brief and reading the guides costs nothing. Paid motion transfer is priced at 1 credit per second for 720p or 2 credits per second for 1080p. Credit packs cost $4.99 for 20 credits, $12.99 for 80 credits and $24.99 for 200 credits. No free generation is offered. One-time credit checkout is available.
Selecting references and saving a text brief stay in your browser. Generate uploads your prepared photo and the selected motion template to private site storage and Kie for AI processing. Completed and failed task files are scheduled for site removal after 30 days; unresolved requests stay for review. See Privacy for details.
No. Bulin 47 AI is an independent project, with no affiliation or endorsement by Bulin 47 or Kevin Wai. Use your own or authorized characters and make the AI origin clear when sharing a result.
The next performer could be you