Create a Rap Selfie Video from One Photo

One photo. Your rap performance.

Free 5-second preview → Unlock full video

Available model

One model is currently available. Source models may differ.

Custom prompt / creative direction

Create a music video; the images are exclusive appearance references and @[blackout video] is the editorial clock. Keep four distinct members, continuous choreography, and an open rightward exit.

Use @[image1] only for Vocal 1's final identity and wardrobe.
Use @[image2] only for Vocal 2's final identity and wardrobe.
Use @[image3] only for Vocal 3's final identity and wardrobe.
Use @[image4] only for Vocal 4's final identity and wardrobe.
Use @[blackout video]  only for vocal order, lip sync, hits and mix; add no words or voices.
Only four performers exist: one instance each of @[image1], @[image2], @[image3] and @[image4]; no doubles, extras, reflections, shadow figures, partial bodies or spare limbs.

LOOK + WORLD: Textured hand-painted 2D anime K-pop video with silhouettes, paper grain. Cycle through a crimson tunnel, cream-coral room, blue-gray rooftop, silver workshop and matte-black moon stage; nonreflective floors, red, cream and cyan light, wide-lens depth, mist.
ACTION + CAMERA: Sharp weighted idol choreography locks to 148 BPM subdivisions. Cut on action with pose-matched set swaps and one stutter edit; rise from solo snaps through a three-cut burst into the chorus, then carry rightward without a final pose.
VFX SYSTEM: Red-cyan light smears, tape-bloom halos and flat geometric flash panels appear only in P06, P07 and P11, never covering faces or lip sync.
SOUND: Use @[blackout video]  unchanged: four distinct female vocals, syncopated snares, 16th-note hi-hats, synth stabs, sub-bass drops and frame-drum pulse; preserve tape saturation and plate reverb.

BEATS:
P01: Low front 3/4 WS, track: exactly four performers; Vocal 1 leads in the crimson tunnel with cap taps and elbow punches while the other three dance behind her.
P02: Dutch profile MCU, push: Vocal 2 dances alone in the cream-coral room through alternating shoulder switches.
P03: High front 3/4 FS, crane: Vocal 3 dances alone on the rooftop and lands a fast heel-toe chain.
P04: Ground frontal MLS, tilt: Vocal 4 dances alone in the silver workshop and drops into a wide weighted step.
P05: Frontal WS, locked: exactly four performers form one diamond in the crimson tunnel and load crossed forearms on the selected snare.
P06: High front 3/4 WS, pull: exactly the same four burst-cut to the cream-coral room, open their elbows and widen the diamond as coral-white panels flash, then vanish.
P07: Low frontal WS, push: exactly the same four burst-cut to the rooftop and release into a low stance; red-cyan smears rise from four floor contacts, then dissolve.
P08: Rear 3/4 WS, track: exactly the same four cut to the silver workshop; three torso hits break the diamond and each member enters her assigned lane.
P09: Top-down WS, crane: exactly four complete bodies kick-cut to the matte-black moon stage, cross in one X, two front and two rear, then open into one line.
P10: Dutch side FS, track: exactly four members travel screen right through the chorus hook; the only stutter repeats one motion slice, never people or body parts.
P11: Low front 3/4 WS, push: exactly four members connect an arm wave, chest hit and stomp beneath the full moon; its tape-bloom halo swells clear of faces, then settles.
P12: Front 3/4 WS, pull: exactly the same four turn clockwise and dance screen right as the world cools into a moon-blue horizon; cut in motion, no final pose.

3367/20000 characters. One photo required. A custom direction replaces the template scene; safety and format rules still apply. Lyrics, language, lip-sync and beat timing are requests, not guarantees. Audio and multi-reference inputs are not supported.

Community source: @aimikoda · github

Your photoOne subject, facing the camera. JPG, PNG or WebP, up to 5 MB.JPG · PNG · WebP

Choose a rap template

Full video length

Upload one photo to enable the free preview.

Street Freestyle scene illustration — Scene illustration. Watch the Street Freestyle sample.

Local static preview only. It does not demonstrate rap motion, vocals or final output quality. Photo checks and previews stay in your browser.

By the Rap Selfie teamUpdated Oct 24, 2026

How does the photo check work?

The photo check works in 4 steps, and every step runs in your browser tab.

Add one photo of the performer. JPG, PNG and WebP files of up to 5 MB are accepted.
Read the score for each photo. Each one gets an overall score from 0 to 100, a short verdict and a tip for any check that needs work.
Replace a photo that scored low. A new photo is checked as soon as you add it.
Continue to the free scene preview, where your photo is placed into the orange Rap Selfie studio.

What does each check measure?

The photo check measures 4 things in each photo and gives every photo an overall score from 0 to 100.

Fact	Value	Source
Resolution	Shorter side of 800 px or more is Good; 500 to 799 px is a warning; under 500 px is flagged	Built-in check
Brightness	Overall light level and left-to-right balance, from Too dark to Too bright	Built-in check
Contrast	Washed-out areas, crushed shadows and flat contrast	Built-in check
Sharpness	Fine detail in the best-focused regions: Sharp, Soft or Blurry	Built-in check
Overall score	0 to 100 for each photo, with a verdict and tips	Built-in check
Faces	Not measured: use 1 person per photo, facing the camera	Built-in check

Questions about the photo check

Which photo files can I use?

Use JPG, PNG or WebP files of up to 5 MB each. Pick the original photo rather than a screenshot or a chat-app copy, because the resolution check reads the pixels in the file you add.

Is my photo uploaded during the check?

No. The photo check and the scene preview run in your browser tab without uploading files. If you leave to pay, your photos save locally in your browser to restore within 1 hour and clear when you return; any draft not restored clears on your next visit.

Why did my photo score low?

A low score points to one of the 4 checks: a shorter side under 500 px, a photo that is too dark or too bright, washed-out or flat contrast, or soft focus. Each check shows a tip, so replace the photo with a sharper original taken in even light and it is checked on the spot.

Do I need an account for the photo check?

No. The photo check and the scene preview are free and need no account. You can buy a video pack and generate directly, too.

What is free and what uses credits?

The photo check and scene preview are free in your browser. Video packs are $9.90 for 1 video, $19.90 for 3 videos, or $29.90 for 5 videos. Each video means 10 seconds; 5- and 15-second videos use the balance proportionally.

Where do the lyrics I type go?

Your words travel from our website through our routing partner kie.ai to the AI video model when you start rendering your video. We keep them in our storage with your video job until it expires after 7 days, or until you choose Delete my data in your account. Your text never appears in emails, and we leave it out of analytics tools.