Make a Gangsta's Paradise AI video from two photos: you and a friend take the rapper and the chorus singer, and the clip's cuts, choir and sound stay.
PNG / JPG / WebP · up to 10 MB · at least 512px on the short side
0/2 photos
Photo 1 is the dark-robe rapper, photo 2 the light-robe singer who takes the chorus. One person per photo, face clear.
Video settings
No prompt needed. Only upload photos of yourself or people who agreed to be featured; celebrity photos and film stills aren't allowed.
The Gangsta's Paradise AI trend is a two-photo video recast of one 30-second clip: you and a friend take its two lead parts, the dark-robe rapper at the front mic and the light-robe singer who takes the chorus, while the hangar, the white-armored choir, the cuts and the sound stay the clip's. It is not an AI cover or a new recording of the song.
It lands because the setting is played completely straight: a vast Star Wars hangar, two robed leads, a rank of white-armored singers, and a performance delivered as if it were a stadium show. The caption on the post it spread from is the joke in one line — “The Jedi Council was NOT ready for this” — and the mismatch between the armour and the chorus is the whole gag.
The choir is why “gangstas paradise choir meme” is a search of its own: the white-armored singers answer the lead line on the beat, so the video reads as a full performance with a backing section rather than a character with a track over the top. Those singers are the clip's, and no photo replaces them.
It is also cut as a performance with a handoff in it: the rapper carries the first half, points at the lens, and the light robe steps up for the closing chorus in front of the same back row. A recast has to keep that timing, which is why this page asks for two photos instead of one.
The hangar clip is posted on Instagram by @alik.bilyalov and @syntx_global, tagged as AI, and reads 442.4K likes and 4.2K comments by 10 November.
The same account posts a second one with a different film pair, read at 100.9K likes and 1.2K comments: the same format with new leads, not a new template.
Template pages for the effect appear within days, each with its own slots, lengths and promises.
Likes and comments read on the Instagram posts on 10 November 2026. A TikTok repost of the same clip was read at 3,099 plays the same day, so the two numbers do not measure the same thing.
Six stills from the cut, each stamped with the second it starts. The rap duo swap who sings once, and the choir is behind them the whole way.
10:01Close on the dark robe at the front mic, half turned, already mid-line. The rank of white armor stands behind him in the same cool light.
20:07The camera turns to his profile and holds it through the next lines while he raps across the mic. A profile is the hardest frame for a recast: there is no second eye to check the likeness against.
30:13The wide shot: the rapper holds the mic on the left, the light robe walks in from the right and stands beside him without singing yet. Both leads share the frame from here on, so the two faces have to stay separate inside one shot.
40:17The cut comes back tight and he aims an index finger straight down the lens while still rapping. His sleeve crosses the frame on the way in, which is the one moment where a recast can smear a hand.
50:24The camera drops to the rank of white armor: helmets down the line, mouths moving together on the answer, and a head in soft focus at the bottom of the frame. This is the shot that makes the clip read as a full performance.
60:28The closing seconds: the light-robe singer comes forward with his own mic, the rapper turns to watch him, and the choir keeps answering behind both. Every recast has to land this shot.
The same template and the same two photos; the only difference is which part of the take you get. The short cut starts at the halfway mark, where the first lead is already at the lens.
Full take · 30s
Opens on the close-up at the front mic, moves through the side shots and the point at the lens, and hands the chorus over in the last seconds. This is the default, and the one to pick if you want the whole arc.
Short cut · 15s
Starts on the point-at-camera close-up: you still get the wide hangar shot, the choir and the handoff, but the rapper's first half is gone. Cheaper, and it still ends on the second lead.
The search mixes four different requests, and only the last one puts your own photos into the performance.
| Compare | This recast | AI cover | Blues remix | Song facts |
|---|---|---|---|---|
| What it makes | Two photos into the clip | Re-sung vocals, new visuals | The same song, another genre | Lyrics, meaning, release facts |
| Who is on screen | You and a friend | No faces, only a voice | A singer covering it | The artists, in text |
| Sound | The clip's own track | A new vocal take | A re-recorded backing | Clips at most, no audio |
| What you get | The cut as an MP4 | An audio or video file | An audio file | A reference page |
The two parts, the timing and the sound are already in the clip, so the work is picking the right photo for each part. Photos bring the face and hair; the robes come from the clip.
Photo 1Dark robe
The lead who opens: he raps the first half at the front mic, points at the lens and turns to watch the chorus. His face is in three-quarter view in the side shots.
Photo 2Light robe
He walks in beside the rapper around the wide shot and takes the chorus in the last seconds, so his face carries the closing frames of the clip.
Works well
Face lit, front-on and unblocked
An unfiltered phone photoAvoid
Two people in one photo
Sunglasses, a mask or a hand over the face
Dark, blurry or far-away photos
Pets: this clip has two human leadsPut one photo in each slot
Both slots take a face before the button turns into Generate; half a pair will not run.
Click Generate
Pick the cut and the resolution; the button shows the exact credit price for that combination before anything runs.
Post it
Download the MP4 and post it with the sound on: the clip's own track is half of what people recognise.
Two public prompts for taking the same two photos to another tool: a Kling video-edit prompt that keeps this clip, and a Seedance prompt that re-shoots the scene when you have no clip to attach. Both are written as instructions, so the tool knows which file does what before it starts.
Kling 4.0 video edit
Prompt length 1030
Use @Video1 as the source clip. Replace only the dark-robe rapper at the front mic with @Image1: the man in the brown leather coat and curly hair who opens the take, points at the lens in the second half and then turns to watch the chorus. Replace only the light-robe singer with @Image2: the bearded man in the pale tan robe who walks in beside him at the wide shot and takes the mic for the closing chorus.
Take the face, hair and skin from each photo and keep the clip's wardrobe: the dark leather robe on the first lead, the pale robe and belt on the second. Keep every cut, the camera moves, the 4:3 framing, the hangar, the rank of white-armored singers behind them, the lighting and the pacing exactly as they play. Keep the clip's own sound and do not add music.
Do not replace the choir or put a face from the photos on a helmet. Do not add people, text or logos. Each replacement stays on its own body and stays stable from the first frame to the last, and the two faces must not blend into each other across the cuts.Give the handoff a second
The chorus changes hands in the closing seconds. Say so, or the re-shoot keeps the first lead to the end and the second photo never sings.
Say which file is the clip
@Video1 is the performance and the sound; both photos are @Image1 and @Image2. A prompt that never says which file is which gets one face on both leads.
Name the wardrobe, not just the role
Both leads wear robes and nothing else separates them at a glance. Naming the dark leather coat and the pale robe is what keeps the two replacements apart after a cut.
Don't re-describe the scene
A long description of the hangar invites a re-stage: new angles, new lighting, and the cuts that made the clip disappear.
Don't ask for a new outfit
Clothes from the photo break the one contrast the clip is built on: dark robe in front, pale robe behind.
Don't leave the choir unnamed
Silence about the back row is how a photo's face ends up on a helmet. One line keeps the choir as armour.
A wish, not a prompt
Make the Gangsta's Paradise AI trend with our photos.
A prompt with the files in it
Replace the dark-robe rapper with @Image1 and the light-robe singer with @Image2; keep @Video1's cuts, choir and track.
The second line names both leads, both files and the parts of the clip to protect; the first leaves the model to invent all three.
The wide two-shot is where both leads share the frame, so that is the frame to check first.
Only the hand is affected; the rest of that second is a real part of the shot, not an error.
The back row is sharp in most shots, so give it one clear line in your own prompt, or check the helmets on the first run.
Every template works the same way: add your photos, press Generate and get a shareable video without writing a prompt.
View all effectsFAQ
It runs on credits: a new account starts with 10 of them, and the button shows the exact price for your cut and resolution before anything is charged. A failed run returns its credits.
Two photos are required, one per slot, and nothing is filled in for you. The same person can go in both slots if you want to play both leads, but then the two faces are meant to look alike — which is exactly what makes the wide shot harder.
No, and this page does not make one. An AI cover rebuilds the vocals; here the vocals, the staging and the sound are the template's clip, and the only thing that changes is the two faces in front.
Yes. The clip comes with its track and the recast keeps it, on both cuts. There is no separate track to add and no way to swap it here; the sound is part of what people recognise in the clip.
No. The photo supplies the face and hair; the robes come from the clip, because the dark and light robes are how the two leads stay apart across the cuts. A photo of a jumper or a cap will not carry its outfit into the shot.
No. The full take is the whole of the built-in performance, not the song, and the short cut starts halfway through that same take. Both keep the same two photos and the same sound.
The frame follows the source clip, so it is 4:3 and there is no vertical version. 1080p is the sharpest on offer; a vertical cut would crop the choir out of the wide shot, so it is not an option here.
The clip cuts from close-ups to a wide two-shot, so both faces are on screen together within the first half. Two similar photos — same hair, same light, same angle — give the model less to keep apart, which is why the photo guide asks for two clearly different people. A face that drifts is a failed run, not something a longer prompt fixes.
No. The clip has two human leads in robes, one rapping at a mic and one taking a chorus, so both slots take photos of people.
No prompt is involved: the two slots, the clip and the recipe are fixed, so you add photos, pick a resolution and click Generate. The prompt kit on this page is for taking the same pair of photos to another tool, and it says what that route loses.
Results from this template stay private on every plan, out of the public gallery. The second photo is usually a friend's, so there is no privacy switch to check — it is already private.
No. This is an unofficial, fan-made video template built on a clip that is itself an AI production; it is not made, approved or run by anyone connected to the film, the characters or the recording.
Two photos, one take
Two photos in: the dark-robe rapper at the front mic, the light-robe singer on the chorus, the hangar and its sound untouched.