Bulin 47 AI Photo Rejected? Fix the Size, Format and Framing in Your Browser

The Bulin 47 AI trend only asks you for one thing: a photo. A generator such as bulin47trendai.com takes a single portrait and recasts the viral dembow street freestyle with that person as the performer. No dancing, no filming.

So it is frustrating when the photo is the part that goes wrong. The upload gets rejected, or it goes through and the face in the video looks soft and only half like you. In almost every case the cause is the file, not the AI.

Before you change anything, compare your photo with the upload requirements on the Bulin47 Trend AI help page. At the time of writing they are:

• a still JPG, PNG or WebP image
• at least 512 × 512 pixels
• no larger than 10 MB
• one clearly visible person, who is an adult and agreed to be in the video

Each section below fixes one of those, using free tools on this site that run in your browser, so your photo is never uploaded to us.

Problem 1: The photo is too small

This is the most common one. Photos saved from WhatsApp, Instagram or a screenshot are often compressed to a few hundred pixels. Even when a small photo is accepted, the AI has very little detail to work with, so the face in the video comes out blurry.

Fix: run it through the Image Upscaler.

1. Upload the photo and choose Real-ESRGAN, the model meant for portraits and real-world photos.
2. Use 2x for a photo that is close to 512 pixels, or 4x for a tiny one.
3. Download the result and check that the shorter side is now well above 512 pixels.

An upscaler can sharpen what is there, but it cannot invent your real features. If the original is a blurry group shot from years ago, a new photo will beat any upscale. The upscaler outputs PNG, so if the file is now large, see Problem 2.

Problem 2: The file is over 10 MB

Modern phone cameras and upscaled PNGs easily pass 10 MB.

Fix: use the Image Compressor and bring it down to a few megabytes. A portrait does not need to be huge: anything around 1,500 to 2,500 pixels on the long side is plenty, and at that size a good-quality JPG is usually 1 to 3 MB. Do not crush it to a tiny file size, as heavy compression adds blocky artifacts around the eyes and hair, and those carry into the video.

Problem 3: The format is not accepted

GIFs, TIFF, BMP and AVIF files will not pass a JPG, PNG or WebP check.

Fix: the Image to JPG converter handles PNG, GIF, BMP, TIFF, WebP and AVIF. An animated GIF becomes a single still image (usually its first frame), so check that your face is clear in the result.

iPhone photos in HEIC are the exception, as this converter does not read HEIC. The quickest fix is on the phone itself: share the photo to yourself through email or messages (iOS usually converts it to JPG), or set Settings → Camera → Formats → Most Compatible before taking new photos.

Problem 4: There is more than one person in the photo

Your best photo of yourself is often a group shot. One-photo recast tools need one person, otherwise they may pick the wrong face or blend two.

Fix:

• If the other person is at the edge, crop the image down to just you. Keep your head, shoulders and upper body in frame, and leave a little space above your head.
• If they overlap you, use the Magic Eraser to brush over them and fill the gap. Check your shoulders and hair afterwards. If the erase cuts into you, crop instead or choose another photo.

Only use photos of other people if they agreed, and only if they are adults. Removing a friend from your photo is fine; putting a friend into a video they never agreed to is not.

Should you remove the background first?

No. It is tempting on a background remover site to say otherwise, but the recast replaces your background with the original street scene anyway, so a cutout gains you nothing. Upload the normal photo, with your hair and shoulders intact.

Where background removal does help is after the video. A cutout of you makes a great cover or sticker for the post: run your photo through the Background Remover, drop the cutout onto a bold color with two or three words of text, and use it as your Reels or TikTok cover so your version stands out from all the others using the same clip.

A 60-second pre-upload checklist

Making the video

Once the photo passes, the rest is quick. If you have not done it before, here is how to make a Bulin 47 AI video from one photo, step by step: upload, choose the original soundtrack or a silent video, review the quote and confirm. The result is a 10-second, 720p vertical MP4.

Fixing the photo first also saves money. Checking the quote is free, but generating uses credits, so it is worth reading the Bulin47 Trend AI pricing before you start. Each 10-second video is 50 credits, and packs are one-time purchases from $9.99. A clean photo the first time is cheaper than a second take.

Quick answers

Why does my Bulin 47 AI video not look like me?

Usually the photo: too small, too compressed, side-on, filtered or badly lit. AI also redraws every frame, so hands and lip sync can vary. Start with a sharp, front-facing photo of at least 1,000 pixels.

Is it safe to fix my photo with these tools?

The tools in this guide process images in your browser, so the photo is not uploaded to our servers. Your final upload only goes to the video generator you choose.

Can I use a selfie?

Yes, as long as it is sharp, front-facing and taken at arm's length so your shoulders are visible. Very close selfies distort the face, so hold the phone a little further away or ask someone to take the photo.

Do I need to label the video as AI?

Yes. TikTok, Instagram and YouTube all ask creators to mark realistic AI content, and it keeps the trend fun for everyone.

Bulin 47, Kevin Wai and their labels are not affiliated with Bulin47 Trend AI or BG Remove Pro. Requirements and prices were checked on October 16, 2026 and may change.