AI video effect
One photo in, one 5-second Bulin 47 clip out: The original street freestyle, one photo in. No filming, no editing, no app to install.
Your photo
One photo is enough: solo or the two of you.
Effect
Bulin 47More effects
Video & sound
This video: 4 credits · 10 songs a month for $29.90, or 40 a year for $100 View pricing
What it looks like
The short answer
Bulin 47 is an AI video effect that turns one photo into a 5-second clip: The original street freestyle, one photo in. You upload the photo, pick the effect, and the model animates the person inside it. It costs 4 credits, and a failed render returns them.
The Bulin 47 effect is one preset, not a blank prompt box. The look — framing, lighting, camera move and how the person moves — is already written into it, which is why two people uploading the same kind of photo get the same kind of clip. Your photo still decides the face, the hair and the outfit; nothing about the person is invented.
How it works
A face-on photo where the person is not cut off, at least 300 pixels on the short side. The panel rejects anything the model cannot read before it charges you.
The face in your photo is the face in the video. A blurry or angled shot is the single most common reason a clip looks wrong.
The effect is already picked on this page. Open the list only if you want a different one — the preview above updates to that effect.
Every effect is a different camera and a different performance. Switching after you generate means paying for a second clip.
The clip renders at 5 seconds. Keep the tab open; the progress bar is a real estimate, and the video appears in the player when it lands.
Rendering happens on servers, not your phone. Nothing is installed and nothing is uploaded from your device after the photo.
Every clip, itemised
The Bulin 47 scene is fixed in the prompt: background, light and mood are the same for everyone who uses it. Your photo supplies the person and nothing else — no wardrobe choices, no location scouting, no blocking to figure out. The clip you see is the clip every user of this effect gets, with their own face in it.
Locked off, medium shot, 35mm, shallow depth of field. No pan, no zoom, no cut — one continuous take.
Head nods, weight shifts and small hand gestures timed to the beat. Neither the person nor the framing morphs mid-clip, which is the usual failure mode of AI video: the face drifting between frames. Here the model is asked to hold one face, one outfit and one body shape for the whole take.
5 seconds by default, which is what short-form platforms reward. Longer lengths cost more because pricing is per second.
16:9 landscape by default; switch to 9:16 for TikTok and Reels before you generate.
The clip ships muted as an MP4. Add your own track in CapCut, Reels or the TikTok editor — most people do, because the platform audio library gives you trending sounds for free. There is no voiceover and no generated music, so nothing competes with your track.
Finished clips land in your account, where you can download them again or re-render with a different photo. Nothing is deleted when you generate the next one, and the same photo can be run through a second effect without uploading it twice.
What people make
“A photo of me, a friend, a group”
Post it as the punchline to a story you are already telling on your timeline.
“An old photo of someone you miss”
Scan a printed photo on your phone and upload the scan. The effect handles the rest.
“One photo with both of you”
Anniversaries, long distance, or just a Tuesday. The two-person version puts both people in frame.
“A photo of your dog or cat”
Pets work because the model only needs a clear face. A sitting, front-facing shot reads best.
“A founder photo or a product shot with a face”
Short, loopable, and cheap to make in batches for a launch week.
“A photo of parents or grandparents”
Upload the best photo you have, even if it is old. It needs a face, not a resolution record.
What makes a good result
Profile shots and sunglasses cost you the likeness. The model needs eyes and a mouth it can read.
A busy background or a crowd teaches the model the wrong thing. Plain walls work best.
Avoid harsh side shadows and heavy filters. The effect adds its own lighting on top of yours.
Only upload people who agreed to it. Anything else gets rejected in review, and paid credits are not refunded for a rights complaint.
Questions
What people ask before they spend their first credits on this effect.
Yes. One photo is the standard input for Bulin 47; upload it, keep the effect selected, and generate. A second photo is optional for the two-person version.
Up to 2. One is enough; the panel tells you when a second photo changes the result.
No. It costs 4 credits per clip at the default settings, and you can try the whole flow — upload, choose, preview pricing — without paying.
No. The MP4 ships silent, so you can lay any track under it. Add your own song in CapCut, Reels or the TikTok editor — that is what most people do.
Usually under a minute for a 5-second clip. Busy hours can stretch it; the progress bar estimates the wait and the player updates when it is done.
The credits return to your balance and no clip is charged. You can retry with the same photo or a different one.
No. You download a clean MP4 that you can upload anywhere, including platforms that scan for watermarks.
Yes, for content you have the rights to. You must own or have permission to use the photo and everyone in it.
Ready when you are
One photo, 4 credits, about a minute. If it does not render, the credits come back.
Failed renders are refunded automatically — you never pay for a clip you did not get.
