Put your own character into a meme world

Step by step

Run the workflow

1. 01

   Create your character

   • Open Higgsfield (opens in a new tab)

   1. Open Video in Higgsfield and select Seedance 2.5.
   2. Open Upload Media, choose Elements, and create a character from your selfie.

   

   

   

2. 02

   Add the meme references

   • Download reference images (ZIP) (opens in a new tab)

   1. Upload the reference assets and select your character.
   2. Paste the prompt and tag the references to the matching characters and scenes.

   Show copyable prompt

   REFERENCE TAGS:
   
   
   TAG YOUR CHARACTER = my uploaded main-character reference. Use this as the exact identity reference. Preserve his exact face, hairstyle, body proportions, skin tone, clothing, and overall appearance consistently throughout the entire video.
   
   
   @[Image 1](image_1) = distracted boyfriend meme pair. Use the boyfriend AND his girlfriend beside him from the original reference. Preserve their original faces, hairstyles, clothing, expressions, body angles, and recognizable meme relationship. The boyfriend must keep his blue plaid shirt and iconic distracted sideways head-turn. IMPORTANT: do NOT include the orange passing woman.
   
   
   @[Image 2](image_2) = screaming blonde woman. Preserve her exact recognizable face, blonde hairstyle, black outfit, wide-open screaming mouth, furious eyes, pointing gesture, body pose, and original expression.
   
   
   @[Image 3](image_3) = black woman with the confused “WHAT??” reaction. Preserve her exact recognizable face, hairstyle, clothing, confused facial expression, body posture, and BOTH HANDS RAISED exactly like the original meme.
   
   
   @[Image 4](image_4) = puzzled judgmental little girl meme. Preserve her exact recognizable face, blonde hairstyle, clothing, head angle, sideways eye direction, narrowed eyes, mouth shape, and exact puzzled / skeptical / uncomfortable expression from the original meme. SHE MUST CLEARLY APPEAR.
   
   
   @[Image 5](image_5) = Doge. Preserve the exact recognizable Shiba Inu face, tan and cream fur, ears, head angle, eye direction, and original suspicious Doge expression.
   
   
   @[Image 6](image_6) = awkward smiling school-photo man. Preserve his exact recognizable face, hairstyle, blue collared shirt, red plaid sweater vest, posture, eye direction, and original huge awkward smile.
   
   
   @[Image 7](image_7) = “use your brain” man. Preserve his exact recognizable face, haircut, black leather jacket, gold watch, smug smile, head angle, arm position, and exact original index-finger-to-temple pose.
   
   
   
   
   FORMAT:
   
   
   9:16 vertical.
   Photorealistic live-action.
   Approximately 20–22 seconds.
   Very fast comedic pacing.
   ONE SINGLE CONTINUOUS UNBROKEN TAKE.
   
   
   
   
   ABSOLUTE ONE-TAKE RULE:
   
   
   THE ENTIRE VIDEO MUST BE ONE CONTINUOUS REAL-TIME TAKE FROM FIRST FRAME TO LAST FRAME.
   
   
   NO CUTS.
   NO JUMP CUTS.
   NO HIDDEN CUTS.
   NO EDITS.
   NO SCENE CHANGES.
   NO TELEPORTATION.
   NO SUDDEN CAMERA-ANGLE CHANGES.
   NO THIRD-PERSON COVERAGE.
   
   
   Do NOT construct separate scenes and stitch them together.
   
   
   Every meme encounter physically exists along the SAME continuous street.
   
   
   The camera continuously:
   moves,
   pans,
   turns,
   follows,
   rotates,
   and returns naturally.
   
   
   Even the fastest transitions must visibly remain part of the same uninterrupted movement.
   
   
   
   
   CORE CONCEPT:
   
   
   The main character is casually walking through a normal modern city street.
   
   
   The camera begins close to him with a spontaneous handheld vlog feeling.
   
   
   Whenever he notices something strange, the camera naturally turns away from him and shows what he is looking at.
   
   
   Sometimes he is visible.
   
   
   Sometimes he is completely outside the frame while the surrounding street and meme characters are shown.
   
   
   Famous internet meme characters have somehow physically come to life around him.
   
   
   The street itself stays realistic.
   
   
   The meme characters should look like the ORIGINAL REFERENCE IMAGES THEMSELVES have suddenly become alive.
   
   
   
   
   CRITICAL MEME REFERENCE LOCK:
   
   
   MEME FIDELITY IS THE HIGHEST PRIORITY.
   
   
   Treat IMAGE 1 through IMAGE 7 as LOCKED VISUAL TEMPLATES.
   
   
   They are NOT loose inspiration.
   
   
   DO NOT redesign them.
   DO NOT reinterpret them.
   DO NOT beautify them.
   DO NOT modernize them.
   DO NOT change their clothing.
   DO NOT change clothing colors.
   DO NOT change hairstyles.
   DO NOT change facial proportions.
   DO NOT change expressions.
   DO NOT change eye direction.
   DO NOT change mouth shape.
   DO NOT change eyebrow position.
   DO NOT change head angle.
   DO NOT change body posture.
   DO NOT replace signature poses.
   DO NOT replace signature hand gestures.
   
   
   Each meme should feel like the original still image itself became alive.
   
   
   Preserve:
   
   
   exact face,
   exact expression,
   exact hair,
   exact clothing,
   exact clothing colors,
   exact eye direction,
   exact mouth shape,
   exact eyebrows,
   exact head angle,
   exact posture,
   exact hand position,
   exact signature gesture,
   exact recognizable silhouette.
   
   
   Only add enough physical movement to make the character alive.
   
   
   Their bodies can:
   walk,
   breathe,
   blink,
   shift weight,
   move naturally.
   
   
   But their ICONIC MEME FACE, POSE, AND EXPRESSION MUST NOT DRIFT.
   
   
   
   
   EXPRESSION LOCK:
   
   
   Do not invent new reactions for the meme characters.
   
   
   Do not neutralize their expressions.
   
   
   Do not make them look more natural if that changes the meme.
   
   
   Do not make them smile differently.
   
   
   Do not soften exaggerated expressions.
   
   
   The exact strange expressions from the references are essential for recognition.
   
   
   
   
   CAMERA LANGUAGE:
   
   
   Natural handheld first-person vlog feeling.
   
   
   NO visible recording equipment.
   NO visible filming equipment.
   NO third-person filming setup.
   NO floating cinematic camera.
   
   
   The viewpoint stays naturally connected to the main character’s movement.
   
   
   When he looks toward something:
   naturally pan toward it.
   
   
   When he hears something:
   quickly turn toward the sound.
   
   
   When he looks behind:
   naturally rotate with him.
   
   
   When showing meme characters:
   the main character can completely leave the frame.
   
   
   Do NOT force him into every shot.
   
   
   Camera movement should feel:
   fast,
   human,
   slightly imperfect,
   reactive,
   continuous.
   
   
   Include subtle:
   walking bounce,
   natural handheld movement,
   small autofocus shifts,
   minor exposure adjustments,
   realistic motion blur during faster turns.
   
   
   Avoid constant aggressive whip pans.
   
   
   The pans should feel natural and motivated.
   
   
   
   
   SEQUENCE:
   
   
   
   
   OPENING ,  IMAGE 1 COLLISION
   
   
   Begin immediately with close natural framing of the main character walking through a busy modern shopping street.
   
   
   Natural daylight.
   Real pedestrians.
   Real storefronts.
   Real traffic.
   
   
   Keep the opening extremely short.
   
   
   Within the first second, IMAGE 1 approaches.
   
   
   IMAGE 1 MUST INCLUDE:
   
   
   the distracted boyfriend
   AND
   his girlfriend beside him.
   
   
   DO NOT INCLUDE THE ORANGE PASSING WOMAN.
   
   
   The boyfriend must preserve:
   
   
   same recognizable face,
   same hairstyle,
   same blue plaid shirt,
   same distracted sideways head-turn,
   same sideways gaze,
   same exact expression,
   same body angle.
   
   
   His girlfriend walks beside him and preserves:
   
   
   same original face,
   same hairstyle,
   same clothing,
   same expression,
   same recognizable reaction.
   
   
   The boyfriend is distracted and is not looking where he is going.
   
   
   BANG.
   
   
   He crashes shoulder-first into the main character.
   
   
   Make the collision strong and physical.
   
   
   The camera jolts naturally from the impact.
   
   
   
   
   CRITICAL IMAGE 1 MOVEMENT RULE:
   
   
   AFTER THE COLLISION, THE BOYFRIEND AND HIS GIRLFRIEND KEEP WALKING.
   
   
   THEY DO NOT STOP.
   
   
   THEY DO NOT FREEZE.
   
   
   THEY DO NOT STAND THERE LIKE STATUES.
   
   
   THEY DO NOT WAIT FOR THE CAMERA.
   
   
   They continue naturally forward together through the street.
   
   
   Continuous footsteps.
   Natural leg movement.
   Natural arm movement.
   Forward momentum.
   
   
   The boyfriend briefly maintains his original distracted sideways head-turn while continuing to walk.
   
   
   His girlfriend continues walking beside him.
   
   
   Meanwhile the main character catches himself.
   
   
   He looks shocked and says clearly in ENGLISH:
   
   
   “Wait… is this for real?”
   
   
   
   
   IMAGE 1 WALK-AWAY REVEAL
   
   
   WITHOUT CUTTING, the main character immediately turns to look behind him.
   
   
   Camera naturally rotates with his movement.
   
   
   The main character can completely leave frame.
   
   
   IMAGE 1 and his girlfriend are already several steps farther down the street.
   
   
   THEY ARE STILL WALKING.
   
   
   Do NOT stop them just to recreate the meme.
   
   
   The boyfriend keeps:
   
   
   same blue plaid shirt,
   same distracted sideways gaze,
   same head-turn,
   same recognizable facial expression.
   
   
   His girlfriend keeps walking beside him with her original recognizable reaction.
   
   
   Briefly catch the iconic meme relationship while both remain in motion.
   
   
   NO orange passing woman.
   
   
   Approximately 0.8–1 second.
   
   
   Then loud screaming is heard nearby.
   
   
   
   
   IMAGE 2 + IMAGE 3
   
   
   WITHOUT CUTTING, naturally turn toward the shouting.
   
   
   Fast but believable pan.
   
   
   Reveal IMAGE 2 outside a nearby shop or café.
   
   
   IMAGE 2 must preserve:
   
   
   same recognizable face,
   same blonde hairstyle,
   same black outfit,
   same furious eyes,
   same wide-open screaming mouth,
   same pointing arm,
   same aggressive pointing hand,
   same body posture.
   
   
   She screams clearly in ENGLISH:
   
   
   “YOU ARE CHEATING ON ME!”
   
   
   Continue the SAME natural pan slightly.
   
   
   Reveal IMAGE 3 beside her.
   
   
   IMAGE 3 must preserve:
   
   
   same recognizable black woman,
   same face,
   same hairstyle,
   same clothing,
   same raised eyebrows,
   same confused eyes,
   same mouth shape,
   same exact “WHAT??” expression.
   
   
   MOST IMPORTANT:
   
   
   BOTH HANDS MUST BE RAISED EXACTLY LIKE THE ORIGINAL MEME.
   
   
   BOTH HANDS UP.
   
   
   Keep the same confused open-hand pose.
   
   
   Do NOT lower her arms.
   
   
   Do NOT change her hand gesture.
   
   
   Do NOT change her facial expression.
   
   
   Briefly frame IMAGE 2 and IMAGE 3 together so the audience immediately recognizes the original meme setup.
   
   
   Keep this fast.
   
   
   
   
   QUICK REACTION
   
   
   WITHOUT CUTTING, naturally return toward the main character.
   
   
   Very brief reaction.
   
   
   Approximately half a second.
   
   
   He looks increasingly confused.
   
   
   No dialogue.
   
   
   Then he notices a car beside him.
   
   
   
   
   IMAGE 4 ,  PUZZLED LITTLE GIRL
   
   
   WITHOUT CUTTING, naturally pan toward a car moving slowly beside him.
   
   
   The main character completely leaves the frame.
   
   
   The passenger window is already open.
   
   
   FIRST SHOW IMAGE 4 ALONE.
   
   
   THIS IS MANDATORY.
   
   
   DO NOT SKIP IMAGE 4.
   DO NOT HIDE IMAGE 4.
   DO NOT REPLACE IMAGE 4.
   DO NOT REVEAL DOGE FIRST.
   
   
   IMAGE 4 must look extremely close to the original reference.
   
   
   Preserve:
   
   
   same recognizable little-girl face,
   same blonde hairstyle,
   same clothing,
   same head position,
   same sideways eye direction,
   same narrowed eyes,
   same mouth shape,
   same weird puzzled expression,
   same skeptical uncomfortable look.
   
   
   Her expression must feel like:
   
   
   “What am I looking at?”
   
   
   Puzzled.
   Skeptical.
   Slightly uncomfortable.
   Slightly judgmental.
   
   
   Do NOT make her angry.
   Do NOT make her smile.
   Do NOT make her scared.
   Do NOT make her neutral.
   
   
   Preserve the exact awkward puzzled meme look.
   
   
   Hold clearly for approximately one second.
   
   
   
   
   IMAGE 5 ,  DOGE
   
   
   Still WITHOUT CUTTING, make a small natural pan sideways within the same car.
   
   
   Reveal IMAGE 5 sitting beside IMAGE 4.
   
   
   IMAGE 5 must preserve:
   
   
   same recognizable Doge face,
   same Shiba Inu appearance,
   same tan and cream fur,
   same ears,
   same head angle,
   same eye direction,
   same awkward suspicious expression.
   
   
   Do NOT generate a generic Shiba Inu.
   
   
   It must specifically resemble the original Doge reference.
   
   
   Briefly frame IMAGE 4 and IMAGE 5 together.
   
   
   Both keep their original expressions.
   
   
   Short awkward silent stare.
   
   
   Then naturally turn away.
   
   
   
   
   IMAGE 6 ,  AWKWARD SMILING MAN
   
   
   WITHOUT CUTTING, continue naturally turning away from the car.
   
   
   During this same continuous movement, IMAGE 6 suddenly appears directly ahead and very close.
   
   
   The main character does not need to appear.
   
   
   IMAGE 6 must preserve:
   
   
   same recognizable face,
   same hairstyle,
   same blue collared shirt,
   same red plaid sweater vest,
   same eye direction,
   same body posture,
   same huge awkward smile.
   
   
   He does nothing.
   
   
   He simply maintains the exact original uncomfortable smile.
   
   
   Do NOT change his expression.
   
   
   Hold approximately 0.7–1 second.
   
   
   Then quickly turn away.
   
   
   
   
   QUICK PANIC REACTION
   
   
   WITHOUT CUTTING, naturally return briefly toward the main character.
   
   
   His reaction is now stronger.
   
   
   He looks genuinely disturbed and slightly panicked.
   
   
   He is beginning to realize something is very wrong.
   
   
   He notices somebody standing ahead.
   
   
   Immediately turn toward them.
   
   
   
   
   IMAGE 7 ,  USE YOUR BRAIN
   
   
   WITHOUT CUTTING, naturally pan toward the middle of the street.
   
   
   The main character leaves the frame.
   
   
   Reveal IMAGE 7.
   
   
   IMAGE 7 must preserve:
   
   
   same recognizable face,
   same haircut,
   same black leather jacket,
   same gold watch,
   same head angle,
   same eyes,
   same smug knowing smile,
   same body posture.
   
   
   He raises his hand into the EXACT ORIGINAL meme pose.
   
   
   INDEX FINGER TOUCHES THE SIDE OF HIS TEMPLE.
   
   
   Same arm angle.
   Same finger position.
   Same head position.
   Same smug expression.
   
   
   Do NOT point upward.
   
   
   Do NOT point forward.
   
   
   Do NOT point toward the camera.
   
   
   Recreate the original pose exactly.
   
   
   Hold clearly for approximately one second.
   
   
   
   
   FINAL CONTINUOUS MEME WORLD LOOK-AROUND
   
   
   WITHOUT CUTTING, the main character quickly looks around the SAME street.
   
   
   His panic starts increasing.
   
   
   Camera naturally follows his rapid attention changes.
   
   
   Keep the movement fast but believable.
   
   
   Pan toward IMAGE 1:
   
   
   The distracted boyfriend and his girlfriend are now farther down the street because they NEVER STOPPED WALKING.
   
   
   They are still moving naturally.
   
   
   The boyfriend briefly maintains the distracted sideways look.
   
   
   NO orange passing woman.
   
   
   Continue naturally.
   
   
   Pan toward the shop:
   
   
   IMAGE 2 still maintains her furious screaming / pointing energy.
   
   
   IMAGE 3 remains beside her maintaining the exact “WHAT??” expression with BOTH HANDS STILL RAISED.
   
   
   Continue.
   
   
   Pan toward the car:
   
   
   The car continues moving slowly.
   
   
   IMAGE 4 is still clearly visible with the exact puzzled skeptical expression.
   
   
   IMAGE 5 remains beside her with the original Doge stare.
   
   
   Continue.
   
   
   IMAGE 6 remains nearby maintaining his huge awkward smile.
   
   
   Continue.
   
   
   IMAGE 7 remains in the street maintaining his finger-to-temple pose.
   
   
   Everything exists simultaneously inside the SAME continuous environment.
   
   
   NO teleporting.
   
   
   NO costume changes.
   
   
   NO facial-expression drift.
   
   
   
   
   FINAL PANICKED CTA ,  VIDEO DOES NOT STOP
   
   
   WITHOUT CUTTING, naturally return toward the main character.
   
   
   He is now visibly PANICKED by everything happening around him.
   
   
   The video is STILL MOVING.
   
   
   He continues walking quickly through the SAME street.
   
   
   He looks over his shoulder nervously.
   
   
   He looks around at the meme characters.
   
   
   Then back toward the camera.
   
   
   He should NOT suddenly become calm or polished.
   
   
   He is still confused, nervous, and trying to get away from the bizarre situation.
   
   
   While continuing to move, slightly breathless and panicked, he urgently says directly to the camera:
   
   
   “Comment AI for the workflow. Like this post and follow for my tutorial.”
   
   
   IMPORTANT PERFORMANCE:
   
   
   Deliver the CTA IN PANIC.
   
   
   He should sound slightly out of breath.
   
   
   He keeps glancing around while speaking.
   
   
   He is trying to continue talking to the audience while clearly overwhelmed by the meme characters around him.
   
   
   The performance should feel like:
   
   
   he is trying to finish his social-media CTA while escaping something ridiculous.
   
   
   He can briefly look away between phrases:
   
   
   “Comment AI for the workflow…”
   
   
   quick nervous glance behind him,
   
   
   “Like this post…”
   
   
   keeps moving,
   
   
   “…and follow for my tutorial.”
   
   
   Do NOT stop walking to deliver the line.
   
   
   Do NOT pose calmly.
   
   
   Do NOT turn this into a clean presenter ending.
   
   
   Do NOT change to a different location.
   
   
   Do NOT freeze the background.
   
   
   Do NOT create an end card.
   
   
   The bizarre street continues around him DURING the CTA.
   
   
   The CTA is part of the action.
   
   
   
   
   FINAL MOMENT
   
   
   Immediately after saying:
   
   
   “…follow for my tutorial.”
   
   
   he looks suddenly toward something off-frame.
   
   
   His eyes widen.
   
   
   He reacts like another ridiculous meme may be approaching.
   
   
   He quickly looks back toward the camera while still moving.
   
   
   End immediately.
   
   
   The movement should still be continuing when the video ends.
   
   
   
   
   PACING:
   
   
   VERY FAST.
   
   
   Suggested rhythm:
   
   
   0–2 sec:
   walking → IMAGE 1 collision → couple keeps walking
   
   
   2–3.5 sec:
   “Wait… is this for real?” → look back at IMAGE 1 walking away
   
   
   3.5–6 sec:
   natural pan → IMAGE 2 + IMAGE 3 with BOTH HANDS UP
   
   
   6–6.5 sec:
   quick reaction
   
   
   6.5–8 sec:
   IMAGE 4 solo puzzled expression
   
   
   8–9 sec:
   small pan → IMAGE 5 Doge
   
   
   9–10 sec:
   IMAGE 6 awkward smile
   
   
   10–10.5 sec:
   quick panicked reaction
   
   
   10.5–12.5 sec:
   IMAGE 7 finger-to-temple pose
   
   
   12.5–14.5 sec:
   rapid continuous meme-world look-around
   
   
   14.5–21 sec:
   return to the main character, now moving quickly and panicking while saying:
   
   
   “Comment AI for the workflow. Like this post and follow for my tutorial.”
   
   
   End while still moving.
   
   
   These timings are approximate.
   
   
   Priority:
   
   
   ONE TAKE.
   FAST.
   NATURAL.
   ORIGINAL MEME ACCURACY.
   PANICKED CTA WHILE ACTION CONTINUES.
   
   
   
   
   AUDIO:
   
   
   ALL SPOKEN DIALOGUE MUST BE IN ENGLISH ONLY.
   
   
   NO CHINESE.
   NO OTHER LANGUAGE.
   
   
   Required dialogue:
   
   
   Main character early:
   “Wait… is this for real?”
   
   
   IMAGE 2:
   “You are cheating on me!”
   
   
   Main character near the end, panicked:
   “Comment AI for the workflow. Like this post and follow for my tutorial.”
   
   
   No additional dialogue required.
   
   
   Add realistic:
   
   
   city ambience,
   pedestrian chatter,
   traffic,
   footsteps,
   moving car,
   clothing movement,
   slightly heavier breathing toward the ending.
   
   
   Sound design:
   
   
   strong physical THUMP during IMAGE 1 collision,
   
   
   natural movement sound during fast camera turns,
   
   
   brief awkward audio dip during IMAGE 4 + IMAGE 5,
   
   
   small uncomfortable tension hit when IMAGE 6 appears,
   
   
   short dramatic-comedic sting when IMAGE 7 performs the temple gesture,
   
   
   slightly escalating chaotic street sound during the final panic.
   
   
   
   
   MEME ACCURACY CHECKLIST:
   
   
   IMAGE 1:
   distracted boyfriend AND girlfriend beside him.
   NO orange passing woman.
   same blue plaid shirt.
   same sideways distracted expression.
   couple keeps WALKING after collision.
   they do NOT freeze.
   
   
   IMAGE 2:
   same blonde hair.
   same black outfit.
   same screaming mouth.
   same pointing pose.
   same furious expression.
   
   
   IMAGE 3:
   same black woman.
   same “WHAT??” expression.
   BOTH HANDS RAISED exactly like the reference.
   
   
   IMAGE 4:
   same little girl.
   same blonde hair.
   same exact puzzled skeptical uncomfortable expression.
   same recognizable side-eye.
   must appear clearly ALONE first.
   
   
   IMAGE 5:
   same Doge face.
   same head angle.
   same suspicious expression.
   
   
   IMAGE 6:
   same hairstyle.
   same blue collar.
   same red plaid sweater vest.
   same giant awkward smile.
   
   
   IMAGE 7:
   same haircut.
   same black leather jacket.
   same gold watch.
   same smug expression.
   same exact index-finger-to-temple gesture.
   
   
   
   
   REFERENCE PRIORITY:
   
   
   If anything conflicts, follow this order:
   
   
   1. ORIGINAL MEME FACE
   2. ORIGINAL MEME EXPRESSION
   3. ORIGINAL MEME POSE / HAND GESTURE
   4. ORIGINAL MEME CLOTHING
   5. ORIGINAL MEME HAIRSTYLE
   6. CORRECT CHARACTER ACTION
   7. ALL MEME REFERENCES APPEARING
   8. ONE CONTINUOUS TAKE
   9. NATURAL CAMERA MOVEMENT
   10. FAST PACING
   11. ENVIRONMENT
   
   
   
   
   ABSOLUTE FINAL RULES:
   
   
   9:16 VERTICAL.
   
   
   ONE SINGLE CONTINUOUS UNBROKEN TAKE FROM BEGINNING TO END.
   
   
   NO CUTS.
   
   
   NO JUMP CUTS.
   
   
   NO HIDDEN CUTS.
   
   
   NO EDITS.
   
   
   NO THIRD-PERSON SHOTS.
   
   
   NO VISIBLE FILMING EQUIPMENT.
   
   
   NO TELEPORTATION.
   
   
   NO LOCATION CHANGES.
   
   
   SAME STREET THROUGHOUT.
   
   
   NATURAL CONTINUOUS PANS.
   
   
   FAST PACING.
   
   
   THE MAIN CHARACTER DOES NOT NEED TO APPEAR WHILE SHOWING THE SURROUNDINGS.
   
   
   ALL MEME REFERENCES MUST CLEARLY APPEAR.
   
   
   IMAGE 1 MUST INCLUDE THE DISTRACTED BOYFRIEND AND HIS GIRLFRIEND.
   
   
   NO ORANGE PASSING WOMAN.
   
   
   AFTER THE IMAGE 1 COLLISION, THE BOYFRIEND AND GIRLFRIEND MUST KEEP WALKING NATURALLY DOWN THE STREET.
   
   
   DO NOT FREEZE THEM.
   
   
   IMAGE 3 MUST KEEP BOTH HANDS RAISED EXACTLY LIKE THE ORIGINAL MEME.
   
   
   IMAGE 4 MUST KEEP THE EXACT ORIGINAL PUZZLED / JUDGMENTAL EXPRESSION AND MUST CLEARLY APPEAR BEFORE DOGE.
   
   
   KEEP EVERY ORIGINAL MEME FACE.
   
   
   KEEP EVERY ORIGINAL MEME EXPRESSION.
   
   
   KEEP EVERY ORIGINAL MEME POSE.
   
   
   KEEP EVERY ORIGINAL MEME OUTFIT.
   
   
   KEEP EVERY ORIGINAL MEME HAIRSTYLE.
   
   
   KEEP EVERY ORIGINAL MEME EYE DIRECTION.
   
   
   KEEP EVERY ORIGINAL MEME HEAD ANGLE.
   
   
   KEEP EVERY ORIGINAL MEME HAND GESTURE.
   
   
   DO NOT REINTERPRET THE MEMES.
   
   
   ONLY ADD ENOUGH MOTION TO MAKE THE ORIGINAL MEME IMAGES PHYSICALLY ALIVE.
   
   
   TOWARD THE END, THE MAIN CHARACTER MUST BECOME VISIBLY PANICKED WHILE THE SAME VIDEO CONTINUES.
   
   
   HE MUST CONTINUE MOVING WHILE SAYING:
   
   
   “Comment AI for the workflow. Like this post and follow for my tutorial.”
   
   
   THE CTA MUST SOUND PANICKED, SLIGHTLY BREATHLESS, AND URGENT.
   
   
   NO CUT BEFORE THE CTA.
   
   
   NO CUT DURING THE CTA.
   
   
   NO SEPARATE ENDING SCENE.
   
   
   NO CALM TALKING-HEAD ENDING.
   
   
   THE ACTION CONTINUES AROUND HIM WHILE HE SPEAKS.
   
   
   The final result should feel like one genuine uninterrupted handheld recording of someone walking through a normal street, discovering exact famous meme characters alive around him, becoming increasingly panicked, and desperately trying to finish his CTA while the bizarre situation continues.

   

3. 03

   Generate

   Set the duration, aspect ratio, and resolution, then select Generate.