Your AI studio for content that goes viral | 25B+ views

Try it now 👉
Joined February 2026
The End of the World Cup
41
1,055
93
9,737
389,297
DeepSeek V4.1 Flash vs MiniMax M3.1 Flash Both had the same prompt "build a 3D rocket launch in Three.js" MiniMax: $0.17 DeepSeek: $0.28 DeepSeek really understood the assignment without the need for a complex prompt and for just a few cents more it beats MiniMax What do you think?
1
116
DeepSeek V4.1 Flash vs. Opus 5.5 Opus costs 20x more, but is it really 20x better than Deepseek? What do you think?
1
1
643
Looks like Opus remains the king of vibe coding Opus 5.5 created a colorful and lively world, and physics are on point Astra did a good job but it still looks a bit flat... What do you think?
3
164
When someone snitches on Tony Soprano
122
McLovin getting into the club with his fake ID
1
169
Looks like Opus 5.5 remains the king of jellyness 🍉 The top one is Opus 5.5, physics, colors, transparency, everything is on point While Gemini 4 Argon is still struggling with making decent physics Same prompt, but somehow only Opus knows how to jelly 🍉
3,541
When you're Derry's creepiest clown...
103
Walter White after joining Los Pollos Hermanos
1
162
Homelander after season finale...
2
176
The Boys - Season 5 Episode 9
2
244
When Skylar finds out about Walter...
1
122
influencers are DEAD 2016 Dropshipping 2021 NFTs 2026 AI influencers this might be one of the biggest opportunities to hit $10k/month FAST in 2026 here's exactly how make your first video 👇
1
1
162
First, go to flashloop.ai/skills And select "Dancing Influencers" Ask the agent to give you character ideas and upload a video to recreate. Choose a character and let it generate the video Here's what I got in just a few clicks👇
1
118
This TikTok account is already pulling $10K/Month with these baby drama videos And the crazy part is Claude can make the whole video for you. With single prompt it makes: - The script - Characters - Video generation Here's exactly how to recreate it in a few clicks 👇
1
1
2
244
How @suno actually works
1
3
130
Seems like Steve is having fun in San Andreas xitter.cf/rehan_shei/status/2105…
Every game is moddable now! I packaged everything Claude learned modding Terraria and Age of Empires into universal-modder: skills + tools that let Claude Code mod almost any game you own. Engine recon, decompiling, @fal sprites/3D/audio, testing in the real game, even the showcase video. github.com/rehan-remade/univ…
Readers added context they thought people might want to know
The video compiles unrelated mods by other creators (e.g. Minecraft in Elden Ring by @TobynJacobs; real-time battles in Pokémon Emerald by @germanburgardt), not work from the advertised repo. Only the final Terraria/AoE clips match its examples. github.com/rehan-remade/u… github.com/GBurgardt/poke… x.com/TobynJacobs/st… x.com/rehan_shei/sta…
2
227
Claude Opus 5.5 just added Minecraft's Steve into GTA 🤯 xitter.cf/rehan_shei/status/2105…
Every game is moddable now! I packaged everything Claude learned modding Terraria and Age of Empires into universal-modder: skills + tools that let Claude Code mod almost any game you own. Engine recon, decompiling, @fal sprites/3D/audio, testing in the real game, even the showcase video. github.com/rehan-remade/univ…
Readers added context they thought people might want to know
The video compiles unrelated mods by other creators (e.g. Minecraft in Elden Ring by @TobynJacobs; real-time battles in Pokémon Emerald by @germanburgardt), not work from the advertised repo. Only the final Terraria/AoE clips match its examples. github.com/rehan-remade/u… github.com/GBurgardt/poke… x.com/TobynJacobs/st… x.com/rehan_shei/sta…
3
2
1
52
7,196
Claude Sonnet 5.5 just ONE SHOT this Minecraft clone It's also cheaper and FASTER than Opus 5.5🤯
1
1
3
443
TikTok is going crazy over these videos right now This AI influencer made over 40M+ views on a brand new account (in just 10 days) Here's exactly how to recreate it 👇
2
2
3
1,210
First I asked Claude to give me a detailed character prompt for a similar persona as the reference video Here's the prompt: Photorealistic full-body portrait of an original comedic character: a tall, lanky man in his late 20s with an intensely deadpan, expressionless face, slightly wide-eyed blank stare, mouth closed in a flat line. He has a dramatic, oversized helmet-like hairstyle: glossy auburn hair in a perfectly rounded, puffy 1970s blowout with a razor-straight fringe sitting just above his eyebrows, clearly styled with too much hairspray. He wears thin wire-rimmed oval glasses and has a neatly trimmed, slightly too-thin pencil mustache. Outfit: a vintage 1970s three-piece suit in mustard-yellow corduroy, one size too big, with wide peaked lapels and shoulders that stick out slightly past his own. Underneath, a pale mint-green shirt with an oversized pointed collar, and a wide chocolate-brown knit tie. High-waisted flared trousers in the same mustard corduroy, and polished tan leather loafers with white socks visible. A cheap gold wristwatch on a stretchy metal band. Pose: standing perfectly straight and stiff, arms hanging rigidly at his sides, feet together, facing the camera directly, like a mannequin or a passport photo taken too seriously. Setting: a sunny European city sidewalk at midday, pale apartment buildings, a metal lamp post, a zebra crosswalk, bright blue sky with fluffy white clouds. Shot on a smartphone at chest height, vertical 9:16, natural daylight, sharp focus, realistic skin texture, no filters, no text.
🤖 Made with AI
1
685
Then I asked Claude to write a detailed reverse engineer prompt from the original clip. Seedance 2.5 + reference image + reference song Here's the full prompt: Vertical 9:16 smartphone video, a single continuous 24.6-second take with no cuts, choreographed to the song from the uploaded reference. Use the uploaded reference music ONLY for its audio track and musical timing.. The performer is the character from the provided character reference @image1 . Match their face, hair, body proportions, clothing and accessories exactly, and keep them identical in every frame from extreme close-up to full body. It looks like a casual TikTok dance video filmed on a phone propped on a ledge on a sunny European city sidewalk corner. The camera is completely locked off for the entire clip, with zero camera movement: no pan, no tilt, no zoom, no handheld shake. Framing changes only because the performer walks toward and away from the lens. Location: a wide sidewalk of terracotta and salmon-colored square pavers with a grey tactile-paving strip leading to a white zebra crosswalk behind the performer. A grey metal signpost stands at the left third of frame, with tall streetlamps behind it. Across the road are modern cream and tan apartment blocks with balconies and a rounded corner facade, a low green hedge and a small blue street sign. The sky is bright blue with large white cumulus clouds and fills the upper half of the frame. Camera and lens: phone main lens, about 24–26mm full-frame equivalent, placed about 1.1 m high and tilted slightly upward, so building verticals converge toward the top of the frame. Deep focus: everything is sharp at every distance. Lighting: hard, bright afternoon sunlight with crisp shadows on the pavers, phone HDR, saturated colors, deep blue sky. Performance style: a completely deadpan, expressionless face with a wide-eyed stare into the lens throughout. All hand movements are small, precise, sharp and mechanical, locked to the beat of the music, with light rhythmic knee bounces. "Right hand" means the character's own right hand, which appears on screen left while they face the camera. Choreography, beat by beat: 0.0–0.8 s: Extreme close-up. The character stands about 40 cm from the lens, chest and shoulders filling the frame, top of the head cropped. They stare down into the lens, motionless. 0.8–2.3 s: They walk straight backward to a mark about 3 m from the camera, touching their collar with the right hand around 1.3 s, then letting their arms settle at their sides. They are centered, framed from head to mid-shin. 2.4–3.2 s: Both hands rise to chest height, palms facing the chest, fingers spread, fingertips wiggling. The fingertips turn to point toward each other, and the hands meet at the center of the chest by 3.2 s. 3.3–4.1 s: Both index fingers point forward with the hands stacked at the center of the chest, the fingers making small twirling circles around each other. 4.1–4.5 s: The hands sink to the belly, then drop straight down to the sides. 4.6–5.2 s: Both arms swing stiffly outward to about 35 degrees from the body, palms down and fingers spread like a penguin, while the character takes one step toward the camera. 5.3–5.6 s: The right forearm swings horizontally across the belly while the left arm hangs, and the hands meet at the waist. 5.6–6.2 s: The arms drop loose, the body sways toward screen right, and the character steps back to their mark. 6.3–7.2 s: Fists come up. They crank alternately at chest and belly height, then stack knuckle-to-knuckle in front of the sternum and twist as if wringing a towel. 7.3–10.2 s: A hand-over-hand rope-pull, about two swaps per second. One fist rises to the chin while the other drops to the waist, then they swap, with the torso leaning slightly toward screen left and small head tilts. At 9.0 s both forearms stack horizontally across the chest with elbows high. At 9.3 s the fists stack in front of the chin; at 9.5 s the palms clasp together at the chest. Then the rope-pull resumes until 10.2 s. 10.3 s: A quick knee dip; the whole body drops slightly. 10.5–11.0 s: The hands hover at the chest with palms facing each other, one above the other, then turn flat and roll around each other horizontally. 11.2–11.8 s: Both fists rise to the collar with elbows out, as if gripping lapels. 12.0–12.2 s: A sharp torso twist toward screen left with the head dipping and the right arm swinging down across the body. The torso then snaps back, right arm at the hip. 12.3–12.9 s: The hands roll around each other at waist height, ending crossed at the wrists. 13.0–13.3 s: The character side-steps to the right third of the frame and turns their body about 60 degrees to face screen left. 13.3–18.0 s: An old-timey bare-knuckle boxer stance in 3/4 profile facing screen left. Both fists are raised in front of the chest, left fist forward, rolling in small vertical circles. Short jabs go toward screen left, and at 13.6 s and 16.5 s both arms extend fully toward screen left. The character bobs and ducks, dipping low at 14.8 s and 17.5 s, knees bouncing, head turned so the eyes stay on the lens, lips pursed. Meanwhile a white hatchback crosses the road behind from right to left (14.0–14.7 s), and a white SUV crosses from left to right (15.3–16.5 s). 18.0–18.4 s: The guard drops and the character turns to face the camera. 18.4–19.2 s: They walk toward the camera to about 1.5 m, framed waist-up, with the head slightly tilted and a blank stare. 19.2–19.4 s: They step back to about 2.5 m, arms at their sides. 19.5–20.5 s: Alternating index-finger points at the lens, right hand then left, jabbing forward about twice per second. 20.6–21.3 s: They walk straight toward the lens. 21.3–21.9 s: Chest-up close-up, face filling the upper frame, a slight sway. 22.0–22.4 s: The character shifts slightly toward screen right while swinging the right arm up. 22.4–23.9 s: Held pose. The right upper arm is raised with the elbow pointing screen left at head height, and the forearm bends back so a loose fist rests just above and behind the head. The head tilts toward screen right, the eyes cut sideways to the lens, and the body sways slightly. 23.9–24.6 s: The arm drops and the character leans in and slightly down toward the lens until the frame crops at the nose. The clip ends on pursed lips in the upper part of the frame. Physics: grounded, natural body weight. Knee bounces compress into the pavement, hair sways slightly and settles with head movement, and clothing creases at the elbows and shoulders and shifts with the torso. Movement is quick and rhythmic, never floaty. Image look: real smartphone footage, digitally sharp, phone HDR, natural wide-lens perspective distortion at close range, no film grain, no cinematic color grade, no shallow depth of field. Constraints: no camera movement, no cuts, no slow motion, no extra people, no smiling or exaggerated facial animation, no deviation from the character reference, no copying of the reference video's performer, no morphing of hands or face, no mirrored screen direction.
802