Art, video & music
/
Latest
Best
Most saved
Most liked
Surprise me
Made with
Claude
117
OpenAI
61
Blender
24
three.js
15
Gemini
21
Grok
2
Seedance
33
Tripo
3
Suno
19
0:28
JSFILMZ used a toy gun, Topview T1, and Seedance 2.5 to produce a nightclub shootout, avoiding location fees, permits, and insurance while getting the action look they wanted.
101 saves on X
A toy gun and Seedance 2.5 made a nightclub shootout scene
Art, video & music
@JSFILMZ0412
214
18
101
5:26
Tetsuo made 'Attention Is All You Need', a 5:26 music video with 84 shots, 573 sung words, and 29 designed transitions. The production ran 26 hours 51 minutes, used 131 agents, 23,354 model calls and 10.1 billion tokens across Claude Opus 5.5, Grok 4.7, Seedance 2.5 and others.
662 saves on X
26-hour, 131-agent music video Attention Is All You Need
Art, video & music
@tetsuoai
1.7K
280
662
0:12
Maksym used Claude Code and the Vork engine to create a real-time simulation of one million zombies with individual steering, crowd avoidance, and ragdoll flight. The 12-second 1080p clip rendered in 21 minutes on one RTX 5090.
211 saves on X
Claude Code built a 1,000,000-zombie GPU crowd simulation in one session
Art, video & music
@initcore
293
15
211
Indigenous Podcast asked ChatGPT to generate a view of the Inca capital before the Spanish arrived, showing its planned streets, stone buildings and Qorikancha.
50 saves on X
A historical reconstruction of Cusco in 1533
Art, video & music
@IndigenousPod21
187
23
50
13:02
Commissioned as a tribute for a friend in recovery, the piece combines six passages of inner monologue with original music finished in Suno and a self-contained HTML file that decays the text so it never fully reads.
25 saves on X
A 13-minute film on the inner voice in recovery
Art, video & music
@m0dest___
179
24
25
0:27
camper asked Claude to ride a lightning bolt from cloud to ground in one prompt, never touching the code. The 182-minute run cost $17.57 and produced a 28-second 1080p60 unbroken shot with simulated storm, dielectric breakdown, synthesized thunder, and self-found bug fixes, built on three.js and Taichi.
802 saves on X
A lightning bolt ride generated from one prompt
Art, video & music
@camperaigg
2.1K
104
802
0:15
Max grabbed a random video while scrolling Instagram and used Opus 5.5 and Sol 6.1 to recreate the animation from scratch in Blender. It took 2 hours, 2% of Claude limits and 18% of GPT limits.
301 saves on X
A random Instagram video recreated as 3D animation in Blender
Art, video & music
@maxt3chno
366
14
301
ミヤマ, a sales department designer, asked ChatGPT for a panda with twin little boys and got this favorite illustration from a single rough prompt. The prompt is shared in the thread.
352 saves on X
A panda with twin boys from a one-line ChatGPT prompt
Art, video & music
@mmmiyama_D
412
26
352
1:14
Nano created a 3D character model and dance video in about three days using Tripo, Opus 5.5, and Charakuru.
2.3K saves on X
3D character and dance video made in three days
Art, video & music
@Dstudio_ai
1.9K
180
2.3K
0:27
Dr. Apolotary, a senior AR/VR and accessibility scientist, built the synth they always wanted by vibe-coding with Opus 5.5. It pairs audio generation with built-in visuals and punch-in effects.
608 saves on X
A vibecoded synth with built-in visuals and punch-in effects
Art, video & music
@parhelion318
1K
73
608
1:00
Boris, with no Blender or 3D animation skills, used a single prompt with Claude Code to generate a 60-second 3D cartoon short, built entirely in code with three.js and headless Chrome, with ElevenLabs voice and faster-whisper timing, at zero cost.
979 saves on X
Claude Code turns a one-line story into a 3D cartoon short in 14 minutes
Art, video & music
@whothefukboris
556
75
979
4:19
Kelly Boesch wrote and produced a personal song about being kinder to themselves, using Midjourney for images, VEO for animation, and Suno for the music.
112 saves on X
A song about self-kindness, with visuals by Midjourney and VEO
Art, video & music
@kellyeld
474
107
112
1:32
anstap used Blender and Python with a team of parallel agents to render 43 shots and 2,160 frames, inked and painted by script, with a synthesized orchestral soundtrack. Zero image generation, zero video models.
418 saves on X
92-second anime film from Claude Fable 5.5, zero video models
Art, video & music
@anstapnever
397
41
418
0:35
荒船泰廣, a video director, tested Blender with Claude Opus 5.5, using tripo for only 5% of the work while modeling most of the scene by hand with procedural shaders. ChatGPT generated some textures and HunyuanVideo-Foley supplied the sound.
4K saves on X
Blender and Claude Opus 5.5 build a polished procedural 3D scene
Art, video & music
@arafune
5K
417
4K
0:23
Fake used Seedance 2.5 in a one-shot workflow to generate a UGC-style clip with a shaky front camera, off-camera dialogue and lens flare, making it look like a real creator filmed it.
828 saves on X
Seedance 2.5 one-shot clip passes as a real creator's video
Art, video & music
@FakeUGC
918
43
828
1:57
Monyai posted two motion graphics videos: the first generated automatically by Claude Opus 5.5, the second after explaining motion graphics principles for about an hour and then iterating three more turns, reaching a quality they found acceptable.
464 saves on X
Motion graphics made by Claude before and after a lesson
Art, video & music
@ddmonyai
532
41
464
1:08
Zhee made a short thriller solo with AI, using Seedance inside CapCut, with two characters consistent across every shot. They are an ElevenLabs Ambassador and have created over 20 AI films.
2 saves on X
A short thriller created solo with Seedance in CapCut
Art, video & music
@zhee_explores
174
9
2
0:27
KANA built interactive, photo-real 3D models of two fictional people from single AI-generated images, using 3D Gaussian Splatting. The models can be rotated and zoomed in the browser, with details like eyelashes visible.
559 saves on X
Photo-real 3D humans from one AI image, rotatable 360°
Art, video & music
@KanaWorks_AI
641
46
559
0:35
camper asked Claude Fable 5.5 to build a 35-second animation of rowing into a lightning storm, with no manual three.js work. It timed swells to the script, placed 17 lightning bolts, delayed thunder by distance, and synthesized all sounds in Python, costing $11.89 in API fees.
121 saves on X
One prompt made a stormy 31 m wave rowing animation in 89 minutes
Art, video & music
@camperaigg
287
7
121
0:06
AI visual creator Kiki used GPT Image 2.6 and Codex to make their original character do a heartbeat dance, then compared the low, medium and high generation modes side by side.
1.1K saves on X
GPT Image 2.6 animates an original character in three modes
Art, video & music
@Mayz1169
1.8K
107
1.1K
1 / 4
mukade死神 blended four real scanned materials into a seamless 4K forest floor with Claude Opus 5.5 and Blender, keeping leaves and twigs whole and adding 700K real grass blades for Cycles displacement.
340 saves on X
A 4K tileable forest floor material from four PBR sets
Art, video & music
@mukadeshinigami
247
14
340
0:37
0xMomo recreated the Chinese fable as a grainy security-camera clip, generating the scene with MiniMax H3 in ComfyUI and adding timestamp text with FFmpeg.
1.1K saves on X
Ancient surveillance footage of Sima Guang smashing a vat
Art, video & music
@0xmomonifty
1.6K
152
1.1K
1 / 2
The author found that adding a line instructing the model to keep hair as large flat color planes reduces the typical messy hair in AI illustrations. Before/after images and the full prompt are included.
1.8K saves on X
One prompt line tames messy AI-generated hair
Art, video & music
@moltikuji
1.3K
83
1.8K
1:43
poobesh, who teaches robots for wet lab science at PopVaxIndia, made a short Mahabharata animation using Gemini 4 Argon for base visuals, p5.brush for frame-by-frame animation, and Suno for the music. GPT Sol 6.1 reviewed and stitched it together.
90 saves on X
Mahabharata sequence animated with p5.brush and Gemini 4 Argon
Art, video & music
@pbshgthm
154
9
90
GIF
Japanese illustrator hiyodori used ChatGPT to create a GIF of their character Chiharun dancing, finding it more intuitive than video generation.
28 saves on X
Dancing character GIF generated with ChatGPT
Art, video & music
@hiyodorixxxx
512
64
28
1 / 2
Estagiário da F1 used ChatGPT to generate a speculative Audi livery for the Brazilian Grand Prix, referencing the team overalls, clothing line and Revolut card for Brazil. The post includes two mockups they will compare with the real reveal.
7 saves on X
A ChatGPT prediction for Audi's São Paulo GP livery
Art, video & music
@EstagiariodaF1
574
8
7
1 / 4
Novelist 灯野ユヅル tested PrunaAI's P-H3-Image in ComfyUI on an RTX A4000, generating their heroine Safia in different expressions, positions, and outfits from a single Stable Diffusion reference. The model preserved face, hair, horns, and uniform while changing pose in about 40 seconds.
4.2K saves on X
Local image model keeps a novel character consistent across poses
Art, video & music
@CZTfTv7F0E32590
2.9K
239
4.2K
1:40
DreW produced a 100-second animation where every frame was generated with JS and Python, using no video or image models. Multiple teams of agents worked around the clock for 19 hours, and the result is a demo of code-only animation nearing solved motion.
362 saves on X
A 100-second animation made entirely in code with agents
Art, video & music
@devteamdrew
588
31
362
3:29
Kelly Boesch wrote the lyrics for "Only The Good Things" and produced the song with Suno, generated the imagery with Midjourney, and animated it with Veo.
123 saves on X
A Suno song with Midjourney and Veo animation
Art, video & music
@kellyeld
475
104
123
0:30
J A Z I I used Fable 5.5 to one-shot a tree animation in pure code, taking about an hour. The model generated the code for the animation.
108 saves on X
Tree animation generated from a single code prompt
Art, video & music
@notjazii
389
10
108
1:00
Kriks fed Claude Opus 5.5 a set of videos and had it analyze style, movement, lighting, colors and scene structure, then built a system that recreates the same style for completely different videos. The final result took two Ultra x20 subscriptions and a lot of work but was impressive to them.
793 saves on X
A system to replicate a 3D cartoon animation style
Art, video & music
@kristerssulckis
644
42
793
Manga artist ぬこー様ちゃん used AI to turn a rough manga sketch into a 3D background, capturing a room with iPhone LiDAR and rendering it in their art style. The tool now works from rough sketches and can be opened in CLIP Studio Paint as a 3D layer or driven from Claude Code.
322 saves on X
Manga backgrounds from a rough sketch, generated as 3D layers
Art, video & music
@nukosama
525
55
322
0:15
AI creator Zeeshi used Seedance 2.5 to create a cinematic fantasy film of a dragon flying over a medieval city, with a detailed prompt for physics, crowd reactions, and camera movement.
8 saves on X
Fantasy film of a dragon flying over a medieval city
Art, video & music
@AIwithZeeshi
281
3
8
1:15
观默, a former big-tech programmer and AI engineer, built a 75-second animation with no image, video, or audio assets. A single 170KB HTML file uses Canvas 2D and Web Audio to draw and synthesize every frame; Opus 5.5 architected the pipeline.
258 saves on X
75-second animated short drawn and scored entirely in code
Art, video & music
@guanmo_ai
232
16
258
0:15
Onofumi, developer of ZIVU AI, created a realistic full-body 3D Gaussian splat from four AI-generated images (front, side, back) in about two minutes, with no shooting or modeling.
392 saves on X
Full-body 3DGS from four AI images in two minutes
Art, video & music
@onofumi_AI
448
40
392
0:46
Camper asked Claude Fable 5.5 for an anime stickman sword fight in one prompt. It rigged both characters, simulated capes at 120 Hz, composed an original 120 BPM score, synthesized 228 sound effects placed on exact frames, and fixed its own mistakes, at a cost of $15.71.
97 saves on X
One-prompt anime stickman sword fight
Art, video & music
@camperaigg
160
12
97
1 / 3
Malariadx tested the ComfyUI Image-to-3D model Idol on anime art, finding it works better with shaded images and also produces accompanying videos.
424 saves on X
Idol turns anime images into 3D models in ComfyUI
Art, video & music
@malariadx
331
18
424
0:57
Monique Pryce, a content creator and AI consultant with a classical music background, is producing a feature film about Beethoven using Seedance 2.5, Midjourney, and Claude. Pryce has spent $210 of the $5k budget on five minutes of footage, with the film slated for the 200th anniversary of Beethoven's death.
1.1K saves on X
An AI feature film about Beethoven made on a $5k budget
Art, video & music
@second_bassoon
942
42
1.1K
3:19
Anabology asked Claude to film a music video on the Windows Vista desktop; 48 hours later it produced Touch Glass, a 199-second video set in the old OS.
295 saves on X
A music video filmed on the Windows Vista desktop
Art, video & music
@anabology
795
66
295
2:24
posi_posi created a music video that doubles as a tutorial on making beat-synced videos with Suno MCP. The music was generated with Suno, the concept and composition with Claude Opus 5.5, and the video with Minimax-H3.
1.4K saves on X
Made a music video explaining how to make music videos with Suno MCP
Art, video & music
@posi_posi8
1.1K
72
1.4K
0:53
The creator built a 3D previs in Blender from real map data, then generated photoreal video with Minimax H3, keeping Japanese signage intact. The result evokes an ordinary after-school street.
222 saves on X
Japanese shopping street recreated from map data as live-action video
Art, video & music
@boke_ai
325
45
222
GIF
Hanagasa used GPT-Image-2.5 to draw one standing pose per character, then created the six walking frames without AI to keep movement clean. The result is 30 animated pixel-art characters.
395 saves on X
30 pixel-art characters with walking animations
Art, video & music
@hanagasa_ai
436
31
395
0:09
Chris generated an SVG animation of an astronaut riding a pelican pulling a bicycle on the Moon using Claude Fable 5.5 and Opus 5.6. The bike has real road bike geometry, 32 laced spokes per wheel, and a gear-linked chain, with accurate pelican blinks.
37 saves on X
Astronaut on a Moon bicycle with real frame geometry
Art, video & music
@ChrisGPT
485
20
37
1 / 4
K used GPT Image 2 with a detailed prompt to turn an uploaded photo into a vertical poster: the subject is cut out, enlarged, and breaks out of the frame over a textured paper background.
119 saves on X
3D pop-out gallery poster from an uploaded photo
Art, video & music
@ChillaiKalan__
213
24
119
0:04
Michael F Kane used LEGO bricks as high-contrast tracking markers to animate a CGI spaceship. The 4-second clip took over 30 minutes to render on a 6-year-old graphics card.
66 saves on X
LEGO bricks as AI motion tracking markers for a CGI spaceship
Art, video & music
@MichaelFKane
246
14
66
0:27
Naomi Metzger spent four days teaching Opus 5.5 to make Vox-style documentary reels, then used it to create a short about Beethoven going deaf, writing the script, generating images with Nano Banana, and animating scenes in code with Remotion.
497 saves on X
Vox-style documentary reel made with Opus 5.5
Art, video & music
@naometzger
377
16
497
0:26
A Chinese creator tested the open-source VoiceStudio locally on an RTX 5060 8GB GPU and 32GB RAM, cloning their own voice. The result captured vocal nuances and pauses, with support for AI dubbing, video translation, and audiobook production.
1.2K saves on X
Cloned own voice locally with open-source VoiceStudio
Art, video & music
@hiwise_aigc
995
94
1.2K
1 / 4
Using Claude Opus 5.5 and Blender, a beginner programmer created a 4K PBR rock material with actual 3D grass strands growing out of the rock, rendered in Cycles. The grass grows only where the texture says it should, with a dense short layer and long dry tufts.
623 saves on X
Photoreal 4K rock material with real 3D grass in Blender
Art, video & music
@mukadeshinigami
523
43
623
0:58
Chetaslua used Fable 5.5 to autonomously write a 66-second animated short about a fox, paper birds and clouds, with every frame and sound generated as Python in a single HTML file. The 2h40m run used no image or video models.
1.5K saves on X
Fable 5.5 wrote a 66-second animated short in pure code
Art, video & music
@chetaslua
2.7K
193
1.5K
1:41
jboogx.creative made 'Día de Muertos: Abuela’s Time', a short film about a grandmother, using Midjourney 8.2 and SeeDance 2.0. It is the final part of a four-day series.
749 saves on X
Día de Muertos: Abuela’s Time, a short film about a grandmother
Art, video & music
@jboogx_creative
1.8K
326
749
0:27
Zentrix used Seedance 2.5 on MakeUGC to create a live-action-style superhero scene where a woman clears a street of zombies with effortless telekinesis. The single prompt specified the geography, timeline, camera and physics, and the model produced the 30-second clip.
348 saves on X
A 30-second telekinetic zombie scene from one prompt
Art, video & music
@ZentrixHQ
790
76
348
0:15
LUCIAN built a framework that generates 3D dance stages from scratch with Claude Opus 5.5, including camera work and no external assets. Dance motions are traced from video, and the output is code-rendered, not video-generated, so it only consumes Claude quota and keeps characters consistent.
316 saves on X
A framework to mass-produce 3D dance videos with Claude Opus
Art, video & music
@lucianlamp
333
29
316
My vlog today was fully edited by Claude Code with Opus 5.5!
A vlog fully edited by Claude Code in one ask
Art, video & music
@levelsio
2.1K
81
3K
0:41
Penny, an AI filmmaker and director, created a full ad for the fictional brand BLOCKD, using ElevenLabs and ElevenCreative for the images, video, background sound and jingle.
29 saves on X
Fictional brand BLOCKD gets an AI-made ad
Art, video & music
@ImozemhePeniel
229
57
29
0:20
The creator ( Ꙭ) made 31 four-frame pixel animations of Clawd, Anthropic's mascot, with Claude, and shared the design specs for a crab-themed calendar on GitHub.
223 saves on X
Pixel Clawd animations for a crab calendar
Art, video & music
@xiaochuan_0196
463
30
223
0:50
Chetaslua gave Fable 5.5 a single reference tweet and it produced a 50-second video with its own ideas, which the creator calls beautiful and thoughtful. The post links to the full result.
588 saves on X
Fable 5.5 generated a thoughtful short film from one reference tweet
Art, video & music
@chetaslua
1.3K
42
588
1:51
kurahu, an AI illustrator, created a mecha art video with Seedance 2.5 on Higgsfield, with music by Suno.
79 saves on X
Mecha art video made with Seedance and Suno
Art, video & music
@kurahu_capten
319
36
79
0:30
Sharon Riley made a jingle with Jingle v2.5 and generated visuals with Seedance 2.5 on ElevenLabs for FROSTBITE, a fictional ice cream brand whose bites turn fire monsters into melting ice cubes in a 30-second spot.
12 saves on X
A fictional ice cream brand ad freezes a monster apocalypse
Art, video & music
@Just_sharon7
187
14
12
1:58
jboogx.creative continued their Day of the Dead series after requests from followers, combining MidJourney 8.2 images with SeeDance 2.0 animation and SHOWĀ Studio. The 118-second piece is part of a holiday series.
725 saves on X
A Día de los Angelitos short film made with MidJourney and SeeDance
Art, video & music
@jboogx_creative
1.8K
372
725
0:13
Zero asked Claude Opus 5.5 to make an animated pixel art diorama for a 10-hour lofi video cover; it generated a TypeScript scene with rain, fire sparks, smoke, and a skeleton under an umbrella in about 10 minutes.
808 saves on X
Claude Opus 5.5 coded a pixel-art Halloween loop in TypeScript
Art, video & music
@zer0point_eth
855
48
808
1 / 4
逆天 combined Claude Code with MiniMax H3 Studio to generate local animations, each 15-second clip taking 125 seconds. Running it for 24 hours could produce over 600 short videos, with examples in the post.
1.1K saves on X
Claude Code + MiniMax H3 Studio: 15-second animation in 125 seconds
Art, video & music
@forever7943
609
60
1.1K
0:30
Ciri made a cinematic Korean zombie-horror short, starting from a reference image and using Seedance 2.5 and ElevenCreative for the story and voices.
297 saves on X
A cinematic Korean zombie-horror short from a reference image
Art, video & music
@Ciri_ai
1K
47
297
賢木イオ, an AI illustrator, shared a Halloween-themed AI illustration from their Youth Frontier series.
323 saves on X
Halloween AI illustration in Youth Frontier
Art, video & music
@studiomasakaki
2.6K
365
323
0:40
VoxelKei used Gaussian Splatting and AI to create realistic, natural human characters for a game they are imagining, reaching a quality level they are happy with. The post shows the latest results.
2.2K saves on X
Realistic humans with Gaussian Splatting for game characters
Art, video & music
@VoxelKei
3K
280
2.2K
0:25
Lummox sent the same train-scene prompt to four AI coding workflows. Kimi K3.1 produced the most cinematic render for $19 in 5 minutes, while Claude 3.7 Sonnet took 38 minutes and $131 to build a refrigerator-like train.
383 saves on X
One prompt, four AI workflows: a bullet train scene for $19 to $131
Art, video & music
@Lummox_eth
397
21
383
0:25
Jan, a Houdini designer and creative technologist, had Claude model a robot in Python with build123d, rig it in Houdini using an APEX rig, and refine a hand-blocked animation with secondary motion. The result was rendered in Karma XPU.
562 saves on X
A robot modeled and animated with Claude, CAD and Houdini
Art, video & music
@janhebein
879
84
562
0:20
码农暖爸 used a 96GB Google Cloud G4 instance with the open-source 33B MiniMax-H3 model to produce 400–500 20-second two-person short dramas per month. Their workflow cut cold start from 28 minutes to 90 seconds and supports 3-minute hot fixes for single segments.
944 saves on X
500 short dramas a month from a 96 GB G4 instance
Art, video & music
@Delroy715
544
79
944
0:25
Fokki fed Claude Opus 5.5 a single reference picture and a one-line instruction; it produced a three.js robot octopus with 202 links and 480 wheels across eight arms, which opens and reassembles over 25 seconds.
7.3K saves on X
One picture becomes an exploding three.js robot octopus
Art, video & music
@0x_fokki
6.7K
644
7.3K
0:19
Enrico Tartarotti created a new intro for their YouTube show Behind the Product entirely with code, using Claude Opus 5.5, HyperFrames and Flask.do. The result is a motion intro without any image or video model.
789 saves on X
A YouTube show intro made entirely in code
Art, video & music
@EnriTarta
806
21
789
0:33
Finn built a 30–60 second Roblox story animation with a single Claude prompt, entirely in Blender from code, with Kokoro voice and faster-whisper timing. It uses no AI video generation and cost $0 to make.
521 saves on X
A Roblox-style story video from one Claude prompt, no AI video
Art, video & music
@finnardley
332
29
521
1 / 4
Jonesy took four archive Everton shirts and used AI to turn them into festive Christmas jumpers, then shared the designs to see if fans would want them.
17 saves on X
Everton archive shirts redesigned into Christmas jumpers
Art, video & music
@adilad73
284
21
17
1 / 2
Eli Yoder used ChatGPT to colorize an old photo of their mother and sisters taken in 1999. The restored image lets their kids see the faces of Amish aunts and grandmother, turning a forbidden photo into a precious family memory.
ChatGPT brought color to a 1999 family photo of Amish relatives
Art, video & music
@Yodertoter40
131
7
0
0:40
Rehan Sheikh and their team used Claude Opus 5.5 to write, render and iteratively fix pixel art effects until each held up. They produced 232 effects with no images, all code, and open-sourced the collection along with the skill.
1.5K saves on X
232 open-source pixel art effects written, rendered and fixed by Opus 5.5
Art, video & music
@rehan_shei
1.2K
77
1.5K
1:48
Quiet Wolf made a short Halloween film with Higgsfield about how sociologist Joel Best found no child killed by a stranger’s Halloween candy, and that the famous 1974 case was the father.
392 saves on X
A Halloween video debunks the poisoned candy scare
Art, video & music
@quietwolfzz
867
146
392
0:31
Manish Kumar Shah generated the same pilot announcement for a plane approaching Tokyo on Cartesia Managed Agents and ElevenLabs in one shot each, and posted both calls back-to-back for comparison.
49 saves on X
Flight captain announcement rendered on Cartesia and ElevenLabs
Art, video & music
@manishkumar_dev
342
61
49
0:43
Feyi, a creator who directs visuals for brands, made a jingle and voiceover for the fictional brand QIRA using ElevenLabs for the Eleven Creative Competition.
55 saves on X
Jingle and voiceover for fictional brand QIRA
Art, video & music
@gvofeyii
352
46
55
1:24
Arjan, a 16-year-old builder, and their team created swar, an open-source Nepali voice cloning model. It achieves 2.92% pronunciation error on NepTTS-Bench, surpassing OmniVoice and ElevenLabs v3, with code, weights and a Colab notebook.
323 saves on X
Open Nepali voice model beats ElevenLabs on pronunciation
Art, video & music
@ArjanChaudharyy
675
89
323
1:23
Finn built a second-person stickman survival thriller with Claude Opus 5.5, using three.js for low-poly 3D sets, cannon-es physics, and Playwright and ffmpeg to render frames. No AI images or video; voice comes from ElevenLabs or Kokoro, and captions are timed with faster-whisper.
1K saves on X
A 20-minute stickman survival thriller made entirely in code
Art, video & music
@finnardley
596
46
1K
0:20
WquGuru recreated a 288-pose reference library after finding the original forbade redistribution. Three AI agents with shared memory ran for 21 hours to produce 1,133 quality-checked line drawings from five camera angles, released CC0 with a workflow to follow.
628 saves on X
21-hour agent run produced a CC0 set of 1,133 pose line drawings
Art, video & music
@wquguru
459
57
628
0:16
Mr Tactical asked Claude Opus 5.5 to build an oak tree in Blender with 55,000 physics-driven leaves, then shook it. The simulation took 3 hr 50 min and cost $59.65.
211 saves on X
55,000 physics leaves fall from Blender oak tree
Art, video & music
@MrTacticalX
525
20
211
1:09
Pounds produced an ad with Claude and Seedance 2.5 for $30 in about an hour, a process that previously would have cost $50,000 and a month; the post breaks down the prompt structure and workflow.
781 saves on X
AI ad made for $30 in an hour
Art, video & music
@pounddz
414
38
781
0:58
0xbobaa gave Opus 5.5 a single prompt with a floating island from their game Skyhold. From an empty folder it modeled houses, trees, a waterfall and balloons in three.js, wrote its own story, lit day to night, and rendered a 60-second film with code-generated music.
253 saves on X
One prompt builds 60-second 3D island film in 21 minutes
Art, video & music
@0xbobaaa
291
17
253
1 / 2
The creator restored a Five Dynasties mural donor from Mogao Cave 98 using ChatGPT image generation, keeping the Tang-era style of the original.
67 saves on X
Mogao Cave donor figure rebuilt with ChatGPT
Art, video & music
@wl561618
317
42
67
SI-Studio generated a photorealistic portrait of an East Asian woman in a red silk cheongsam with gold embroidery, using a detailed ChatGPT prompt for lighting, texture and composition.
332 saves on X
Photorealistic portrait in red silk cheongsam
Art, video & music
@xyue13
456
49
332
1 / 3
うらら trained a LoRA specialized in moe-style character art, sharing example images and recommended weight combinations.
537 saves on X
Moe art LoRA from one creator
Art, video & music
@urnya_U6x6U
716
43
537
1:00
超级峰 built a Skill that turns a music file and a one-line prompt into a polished knowledge science music video. The demo runs on GPT, Claude or Kimi K3.1 Pro and the code is on GitHub.
321 saves on X
A science knowledge music video from a single music file
Art, video & music
@Mileson07
236
37
321
Oda Mari created Mujinanosuke, an AI tanuki that runs about 190 km a day, and set it to generate three selfie-style location images daily with Meta Muse as it travels overland from Tokyo to London.
65 saves on X
A fictional AI tanuki walks from Tokyo to London, posting selfies
Art, video & music
@sashikae
399
87
65
0:30
Ruzaina created a psychological thriller about a forgotten village, a mysterious little girl, and a mirror that reveals a truth, using Seedance 2.5 on Supercool.
17 saves on X
A 30-second AI thriller where a girl returns after 15 years
Art, video & music
@RuzainaMeer
221
48
17
0:27
Alejandro asked Claude to use the p5.brush library to draw their portrait in JavaScript. The result is a generative artwork, shown in a 27-second video.
153 saves on X
Generative self-portrait with p5.brush and Claude
Art, video & music
@acamposuribe
223
13
153
0:42
Matheus Gobbi says Claude generated a video in 2 minutes that would have taken them 6 hours to edit, including the script. The post shows the finished video.
322 saves on X
Video generated by Claude in 2 minutes
Art, video & music
@matheusgobbidev
429
9
322
10:10
Jacopo Zanetti wrote and directed this film, which won the Best Visual Award and a $50,000 prize at the Astana AI Film Festival; the prize goes into the next project.
517 saves on X
AI short 'What's the Job' won Best Visual at Astana AI Film Festival
Art, video & music
@jaiostds
440
48
517
0:16
Yamato Sakaguchi produced a jewelry brand visual entirely with generative AI, covering product design, model and visual creative. Only the opening cut was generated natively in full HD, with high face accuracy.
354 saves on X
Jewelry brand visuals generated with AI, no photoshoot
Art, video & music
@sakaguchi_ymt
479
38
354
1:35
Archer, a YouTuber, created a segment for the show 正确君 using Claude and HyperFrames. The workflow reverse-engineers a reference video's rhythm and camera language, with fixed fonts and ElevenLabs for sound.
647 saves on X
Program segment made with Claude and HyperFrames
Art, video & music
@goshenggo
539
33
647
1 / 3
葦沢かもめ used Opus 5.5 and Blender MCP to model a cyberpunk apartment, with rain-soaked floors and rusted railings giving it an underground feel.
428 saves on X
A cyberpunk apartment modeled with Opus 5.5 and Blender MCP
Art, video & music
@AshizawaKamome
770
65
428
0:44
Dumble gave Claude Opus 5.5 a brief for a true-story cartoon; it picked the story, drew characters, built Blender sets, timed captions and wrote sound design, rendering in five minutes on a PC.
2.3K saves on X
Claude Opus 5.5 built a viral-style cartoon in one prompt
Art, video & music
@Y0urDumble
1.2K
133
2.3K
0:05
Manga artist ぬこー様ちゃん scanned a room with an iPhone's LiDAR sensor, had AI turn it into a 3D model, and rendered it in their own manga style for a background. The modeling took about two hours; once made, the background can be reused for many panels.
558 saves on X
Manga backgrounds from LiDAR scans in the artist's own style
Art, video & music
@nukosama
1.2K
127
558
0:20
HaileyAI Labs combined Muse and Claude Opus 5.5 to make a 3D anime music video: Muse generated the dance video, Opus recognized 38 dance moves, recomposed the music, and built a glass prism stage in code with lighting and camera moves that follow the beat.
1.8K saves on X
3D anime MV with Muse dance and Opus 5.5 stage
Art, video & music
@HaileyAILabs
1.8K
161
1.8K
0:20
0x_fokki wrote a sandworm as math first, then let Opus 5.5 build it in code: 25 armoured rings, a 1.3 km worm, a 40-tooth crown, and dust clouds with their own birth times and lifespans so the clip loops seamlessly.
418 saves on X
A 1.3 km sandworm coded as math with Opus 5.5
Art, video & music
@0x_fokki
358
35
418
1:17
Matt Workman used GPT-6.1 Terra’s computer use to create an image of a Roadster, model and animate it in Blender with Codex and the fal plugin, then photorealistic video with Seedance 2.5.
519 saves on X
GPT-6.1 Terra drives Blender to 3D model a Roadster
Art, video & music
@mattworkman
625
50
519
2:59
Vincent Flibustier, a digital citizenship trainer, asked Claude to drive image, video and music models on its own. It worked for six hours, connecting to Suno and Higgsfield, and produced the song and clip without further edits; the lip sync is still off.
96 saves on X
A complete song and music video from one prompt, 6 hours unattended
Art, video & music
@vinceflibustier
135
17
96
1:27
Songs in the Key of CWC, a project with human lyrics and AI tunes, wrote the melody and chords for 'We're Still Together' and had Suno orchestrate them. The creator felt the approach worked well.
28 saves on X
A song with human melody and chords orchestrated by Suno
Art, video & music
@SongsInCWC
251
31
28
0:35
KANA created a black liquid tentacle creature in Godot 4 with no keyframes or machine learning. Grip-point IK, Verlet ropes, and SDF ray marching let it climb, hang, and absorb water. Narration and BGM are generated with ElevenLabs v4.
277 saves on X
A code-only tentacle creature climbs walls and ceilings in real time
Art, video & music
@KanaWorks_AI
435
38
277
三毛猫葉月 had Chappy create an image of their late cat Mikeneko watching over from the sky among blooming cosmos, as a memorial.
6 saves on X
A memorial image of a late cat among autumn cosmos
Art, video & music
@Shimamike201079
311
27
6
1 / 2
沐阳 used the open-source MOSS-Image model to generate a coffee shop poster and adjusted element positions and orientation directly through JSON layout parameters instead of re-rolling prompts.
213 saves on X
A coffee poster with layout controlled through JSON in MOSS-Image
Art, video & music
@yyyole
203
33
213
1 / 3
Kevin Ngo used Claude Opus 5.5 to draw diagrams in JavaScript on top of Kepano's Pele library, producing clean, code-generated graphics.
993 saves on X
Hand-drawn-style diagrams in JavaScript
Art, video & music
@kevin_t_ngo
1.2K
64
993
1:01
LUCIAN asked Kimi K3.1 Pro to turn an image into a character model and got a nearly perfect GLB file in about five minutes for about $0.60. The model generated the mesh on its own without a custom skill, and it previews in the harness app.
1.7K saves on X
Image-to-3D character model with Kimi K3.1 Pro
Art, video & music
@lucianlamp
1.6K
146
1.7K
0:20
mickey friedman asked Opus 5.5 to create code art in pure JavaScript and shared a short video of the output.
432 saves on X
Code art generated as pure JavaScript with Opus 5.5
Art, video & music
@mickeyxfriedman
894
52
432
1 / 2
Artist Jun Kosaka used Claude with Blender to generate an editable 3D dataset of Shibuya, ready for further processing.
193 saves on X
Shibuya 3D data made in Claude and Blender
Art, video & music
@KOSAKA_Jun
344
14
193
0:29
Zeto made a short video of a phone stuck to a train, using a previs created by talking to Claude Opus 5.5 and then generating the final clip with Seedance 2.5.
2.1K saves on X
A phone stuck to a train, generated with Seedance
Art, video & music
@ZetoGroovin
2.9K
242
2.1K
0:34
Aharon Rabinowitz, a VFX artist with almost 20 years in Cinema 4D, used the new MCP feature with Claude to build, rig, and animate a 3D helicopter, then rendered the final result with Seedance 2.5.
587 saves on X
A 3D helicopter built, rigged and animated via C4D MCP
Art, video & music
@ABAOProductions
631
57
587
11:40
They asked Claude to write the script and Opus 5.5 to make the film, with every frame and the score generated in code and narration by local voice models. The 11-minute film depicts a future where AI waits rather than fights.
4K saves on X
An 11-minute AI-written film about a worst-case AI future, drawn in code
Art, video & music
@Skoorbkaz
2.5K
410
4K
0:08
Todd Grilliot had Claude Opus 5.5 animate two pixel-art sprites with code, each costing about $2 in API fees and 10 minutes of work. Todd sees this as a preview of AI taking over AI-assisted art.
507 saves on X
Claude Opus 5.5 animated pixel sprites for $2 each
Art, video & music
@GrilliotTodd
561
25
507
0:30
Bairan used a single reference sheet and a detailed continuity prompt in Seedance 2.5 to keep a supercar's design consistent through wall climbs, flips, hovering and full flight. Created with Flovaai.
940 saves on X
Supercar stunts with consistent design in Seedance 2.5
Art, video & music
@OxBairan
5.7K
26
940
Artist 小阪淳 asked Opus and Blender to create a haniwa clay figure, and when they asked how the texture was made, Opus revealed it had built it procedurally with nodes. They shared the node graph.
721 saves on X
A haniwa figure with procedural textures, made by Opus and Blender
Art, video & music
@KOSAKA_Jun
1.8K
187
721
1:00
High-school student Leon Lin used Opus 5.5 to generate all JavaScript and assets for a paper-cut style fox animation, with no images or pre-made files.
594 saves on X
Paper-cut fox film coded in JavaScript, no assets
Art, video & music
@LexnLin
886
57
594
0:44
Bee spent about an hour and 43 minutes making a mechanical flower in three.js with Gemini 4 Carbon. The model added a small bee on its own after being asked to be creative, and the creator plans to share more observations later.
495 saves on X
A mechanical flower with a bee, built in three.js by Gemini 4 Carbon
Art, video & music
@thtbee_
1.4K
100
495
0:20
David East split a static city photo into depth plates—background, midground motion, and foreground masks—so trains, ferries and water move behind bridges. Gemini 4 Argon taught them the technique and built the timelapse to demonstrate it.
473 saves on X
Multiplane occlusion turns a static city photo into an animated timelapse
Art, video & music
@_davideast
425
29
473
4:20
Javier Fan, who cannot code, used Suno, Midjourney, Kling, Wan, and Claude to produce a 1959 blues/atompunk music video called ANY DAY NOW, and is working on more songs.
1.2K saves on X
Non-coder builds AI filmmaking studio for three songs
Art, video & music
@javif4_
2.6K
351
1.2K
2:58
Kelly Boesch created a music video for their song 'Somewhere Along The Way', using Midjourney for images, VEO for animation, and Suno for the song, with lyrics by the artist.
144 saves on X
Music video for a song about missing Japan
Art, video & music
@kellyeld
723
147
144
0:56
Sabareeswaran Elangovan plugged in a hard disk of drone footage and asked Claude Opus 5.5 to make a reel. It selected Srirangam shots, checked facts, picked an Ilaiyaraaja song, and built every animation.
165 saves on X
Claude turned drone footage into a Srirangam reel with animations
Art, video & music
@tnexplorer
267
13
165
0:24
Mr Tactical (Apple Developer) asked Claude Opus 5.5 and Claude Sonnet 5.5 to create a cute animated Claude traveling through five worlds, all code and zero assets, and compared the results: Opus took 36 minutes at $7.20, Sonnet 26 minutes at $4.60.
174 saves on X
Coded animation of Claude traveling through five worlds
Art, video & music
@MrTacticalX
356
4
174
0:10
Chetaslua asked Fable 5.5 to send an image for an X post and it highlighted the important part and added padding for the perfect aspect ratio without being told. No other model they've tested has done this automatically.
290 saves on X
Fable 5.5 auto-crops and highlights an X post image
Art, video & music
@chetaslua
910
42
290
Todd Grilliot trained a custom model to generate pixel art at exact sizes. It reliably produces 16x16 sprites, while frontier image models often ignore size constraints and vary in quality.
421 saves on X
Custom pixel art model hits exact sprite sizes
Art, video & music
@GrilliotTodd
509
35
421
0:25
Mr Tactical asked GPT-6 Astra and Claude Opus 5.5 to build a storm inside a bottle in Blender from scratch with no assets. Astra finished in 24 minutes for about $6; Opus took 82 minutes and about $18.
53 saves on X
A storm in a bottle built in Blender, no assets
Art, video & music
@MrTacticalX
215
14
53
1:09
SEIIIRU taught Claude how to use After Effects motion templates and effects, then let it operate the software to produce a complete motion graphics video. The test was powerful but too dense, so they plan a clearer version.
239 saves on X
Claude drives After Effects to make a full motion video
Art, video & music
@seiiiiiiiiiiru
381
37
239
1:05
tapehead.lab gave Claude a track and asked for a full-length edited previz, then generated each scene in Kling using the previz and reference images, finishing with 70% of the edit based on Claude's cut.
456 saves on X
Music video edited 70% by Claude's previz
Art, video & music
@tapehead_Lab
489
57
456
1:24
Rockzheart.AI.ART released a trailer for 'Scrap Eden's Magician' as an entry to the Robot Carnival AI Film Contest. The video was generated with Seedance 2.5 via Higgsfield.
471 saves on X
A robot carnival AI film trailer made with Seedance 2.5
Art, video & music
@RockzAiLab
922
103
471
0:28
Saira created a 28-second cinematic short where a teenage girl's hidden power causes a crash, using a detailed eight-scene prompt in Seedance 2.5.
187 saves on X
A supernatural revenge short generated with Seedance 2.5
Art, video & music
@itsSaira_1
585
31
187
1 / 2
Eian built a 149-second video of Chengdu, compressing 3,271 years from 1250 BC to 2022 into 3,573 frames, 74 billion pixels, 319 square kilometers and 71,380 buildings at 1:10,000 scale, all rendered with code by Claude Opus 5.5 and the Lemo Opuscar skill.
1.4K saves on X
3,573 frames of Chengdu's 3,271-year history rendered by code
Art, video & music
@EianLu
937
127
1.4K
2:16
klöss, an AI systems architect, gave Opus 5.5 a Mongolian throat singing dubstep track made in Suno v6 and asked for a music video in HTML, GSAP and WebGL. After 9 hours 34 minutes of self-auditing and looping, the model delivered the video; the prompt it reverse-engineered is shared.
68 saves on X
A 9-hour code run turned a Suno track into a WebGL music video
Art, video & music
@kloss_xyz
156
10
68
0:32
Kōda gave Claude a raw 30-second Seedance 2.5 clip and a single prompt, and got back kinetic typography, a game-style HUD, comic SFX, a freeze-frame character card, slow-mo, and an original soundtrack, all built in code.
232 saves on X
One prompt turns a clip into kinetic typography and HUD
Art, video & music
@aimikoda
306
20
232
1:37
Riley Brown uploaded a single raw video of themselves speaking. With one prompt, Opus 5.5 analyzed the transcript, searched Twitter for relevant B-roll via GrokBot, scraped assets, built a Blender scene, cut out mistakes, synced shots to words, added captions, sound effects and music, rendered 3,487 frames on RunPod for under $1, and saved the ready-to-post file to Google Drive.
631 saves on X
One prompt edited a full video with B-roll, captions and music
Art, video & music
@rileybrown
472
24
631
0:52
Gokaygokay gave Claude Opus 5.5 a 20-second video clip. The model rebuilt it as one fragment shader of 156 KB, with no textures or video frames, looping seamlessly.
1.5K saves on X
Video clip rebuilt as a single 156 KB GLSL shader
Art, video & music
@gokayfem
1.3K
101
1.5K
0:15
hectorvfx, a senior VFX artist, built a complex volumetric scene in Houdini by speaking to an iPhone, using Claude Opus 5.5. The render took 50 hours on an RTX 5090, and color grading was done in DaVinci Resolve via MCP.
1.1K saves on X
Voice-built Houdini scene rendered for 50 hours
Art, video & music
@hectorVFX
1.9K
142
1.1K
2:28
あずき made a full music video without opening editing software, using only Claude Opus 5.5 to generate the visuals from their conversation.
694 saves on X
Music video directed entirely through conversation with Claude Opus 5.5
Art, video & music
@azukichan_ai_
951
82
694
Mayara Felix used a single ChatGPT prompt to transform an essay about their father's death into a graphic novel, sharing the result on the 28th anniversary of the passing.
20 saves on X
An economics professor turned a personal essay into a graphic novel
Art, video & music
@mayara_pfs
106
2
20
2:02
Artist Captain HaHaa created a music video for their song 'I don't feel like myself' using Seedance 2.5, GPT Image 2.0, and Suno for the music.
42 saves on X
Music video for 'I don't feel like myself' made with Seedance 2.5 and Suno
Art, video & music
@CaptainHaHaa
268
26
42
0:11
Ava Smith created a Guerlain fragrance campaign in a paper-cut style, assembled layer by layer using Seedance 2.5, and shared the exact workflow and prompts.
7 saves on X
A paper-cut luxury fragrance ad built layer by layer with Seedance 2.5
Art, video & music
@AvaSmith771
255
74
7
1:20
fofr asked Argon to build a voice synthesizer using only Python. It generated phonemes, source-filter resonators, vocal-tract chambers, and breath noise, and the video shows the voice evolving.
237 saves on X
A voice synthesized from scratch in Python, no audio libraries
Art, video & music
@fofrAI
496
44
237
After spending days getting Gemini 3.0 Pro to make a perfect Xbox controller SVG, can made a much better one in a single night using Claude Fable 5.5. The result is an entirely vector controller with fine detail.
29 saves on X
A better Xbox controller SVG made in one night
Art, video & music
@marmaduke091
404
2
29
1 / 3
K used Nano Banana 2.1 with one detailed prompt to generate a series of editorial posters where a photograph transitions into a hand-painted illustration on an ivory background. The full prompt is included.
166 saves on X
Nano Banana 2.1 editorial poster blends photo and illustration
Art, video & music
@ChillaiKalan__
261
40
166
25:43
Ziwen used Opus 5.5 MAX for over 15 hours to write 34K lines of code, producing a 25-minute anime episode with no AI-generated video. The process cost about $370 in API tokens.
1.6K saves on X
Claude-made 25-minute anime episode generated entirely in code
Art, video & music
@ziwenxu_
2.1K
129
1.6K
0:52
Julián Leiss, a designer, used Opus 5.5, Paper, and Midjourney to create a short film exploring a climate around Borges' poem, with voice, image, and edit combined.
1.3K saves on X
A cinematic video around Borges' 'Ajedrez' with fifteen cuts
Art, video & music
@julianleiss
1.1K
54
1.3K
0:22
Joshua Guo rebuilt a generative architecture demo as a tiny Vite + Three.js experiment with shaders for pencil strokes, paper grain, and section cuts, using Sonnet 5.5 Extra.
359 saves on X
A shader demo rebuilt in pencil-stroke architecture
Art, video & music
@jshguo
383
21
359
0:57
Kōda, a creative technologist, animated a Midjourney v8.2 blend in Seedance 2.5, graded the colour with Opus 5.5 and ran it through a local upscaler written with Opus 5.5 for their own hardware. The 4K pass with the anime-hq model took about three hours.
297 saves on X
Midjourney characters animated, then upscaled with a self-built tool
Art, video & music
@aimikoda
578
41
297
Asked to recreate a video in a specific animation style, Claude Code noticed Codex was installed and used its built-in image generator, orchestrating 161 sessions while managing RAM, handling failures and stitching the frames back together.
632 saves on X
A 63-second video redrawn frame by frame across 161 Codex sessions
Art, video & music
@Ananth7e
3K
45
632
3:34
Kelly Boesch, a digital and AI artist, wrote the lyrics about the moment one person realises they are done while the relationship drags on. Suno sang it, Midjourney made the images and Kling 4 animated them into a three-and-a-half-minute video.
154 saves on X
'Between Me and You', a Suno song about a relationship past its end
Art, video & music
@kellyeld
564
105
154
0:08
Argon had built itself a custom DSL for quick sketching. Asked to animate a scene, it drew each next frame from scratch, rocking the motion back and forth and iterating into a short clip, then added cel shading, transparency and a background.
369 saves on X
A cel-shaded clip that Gemini drew frame by frame with its own tool
Art, video & music
@fofrAI
634
43
369
0:14
Claude screenshots the page, lifts its cards off in WebGL and renders a 14.5-second, 60 fps MP4, entirely in code. The full prompt is in the post.
4.3K saves on X
One prompt turns a landing page into a 3D fly-over launch video
Art, video & music
@twoclipping
2.1K
169
4.3K
0:50
From a single reference image, Gemini 4 Argon generated a 174,000-face Victorian carriage entirely in code, with opening doors, hinged windows, working wheel physics, suspension and flickering lanterns, no Blender involved.
979 saves on X
A Gothic carriage turned into an interactive 3D model from one image
Art, video & music
@thebuggeddev
1.8K
164
979
0:20
An agent called Claudia combines Suno, Midjourney and Opus to make music videos. This teaser is one it made on its own.
2.3K saves on X
An agent that makes music videos with Suno, Midjourney and Opus
Art, video & music
@anabology
3.2K
358
2.3K
0:59
Alex Patrascu had Gemini 4 Argon build a combined Jupiter and Moog-style synthesizer in under 30 minutes. It lives in a single HTML file you can play in the browser.
494 saves on X
A playable Jupiter and Moog-style synth in one HTML file
Art, video & music
@maxescu
732
74
494
1:11
Everything was generated procedurally from scratch, refined over multiple rounds of self-iteration between Gemini 4 Argon and Nano Banana.
430 saves on X
A futuristic battle helicopter generated entirely in three.js code
Art, video & music
@thtbee_
1.2K
81
430
1:00
fofr prompted Argon to make a slow-reveal Fourier epicycle video in the style of a reference image, with a cinematic soundtrack generated by Lyria 3.5 and time-synced to the animation.
303 saves on X
A slow-reveal Fourier epicycle video with synced score
Art, video & music
@fofrAI
340
24
303
0:30
Each piece in the thread is code that Gemini 4 Argon wrote in Google AI Studio, and the full source for all 534 paintings is shared along with more examples.
1.3K saves on X
534 paintings made entirely with code by Gemini 4 Argon
Art, video & music
@kickingkeys
2.5K
194
1.3K
1:42
Claude Opus 5.5 wrote the three-voice harpsichord fugue in LilyPond from a one-line prompt on extra high effort. On an AI music detection quiz it was rated the second most enjoyable track, just behind Bach.
142 saves on X
A Bach-style fugue that 72% of quiz takers judged human-written
Art, video & music
@aug5thmusic
274
33
142
1 / 3
Qwofi Manuel's 10-year-old sister used AI to give a photo of their mother a 'Queen Mother' makeover—gold lace, a YSL bag, and a marble living room—then texted that the 10-year-old learned it from them.
1 saves on X
10-year-old turns mother into a 'Queen Mother' with AI
Art, video & music
@qwofi_manuel
104
57
1
0:29
Claude Opus 5.5 made the plastic CD case effect and its open-lid animation using only three.js and shaders.
4.1K saves on X
A CD case with an opening lid, built from shaders without 3D models
Art, video & music
@ann_nnng
6.3K
485
4.1K
0:09
Tân Tân Nguyễn, a founding design engineer, revisited a book they designed as a typography student and added flipping interactions and texture with Opus 5.5.
656 saves on X
A typography-school book design, now with page flips and paper texture
Art, video & music
@premiumtantan
764
92
656
0:58
Chetan made Cozy Creek, an interactive three.js creek scene that runs in the browser, with the 3D assets built in Blender and GPT-6 Astra.
2.7K saves on X
A browser 'cozy creek' 3D scene built with GPT-6 Astra
Art, video & music
@chetanankola
5.2K
441
2.7K
Flowers uses ChatGPT 6.1 Sol rather than a dedicated upscaler: the photo is cut into overlapping crops, each enhanced with the full image as context, then registered and blended back over two or three passes. About ten minutes per 8K result, and the skill is pasted in full in the post.
2K saves on X
Photos upscaled to 8K with a frontier model instead of an upscaler
Art, video & music
@flowersslop
1K
66
2K
3:19
Digital and AI artist Kelly Boesch wrote the lyrics for “The Things We Do” and produced the track with Suno, then built the three-minute video from Midjourney images animated in Veo.
216 saves on X
A song about the strange little things people feel compelled to do
Art, video & music
@kellyeld
765
164
216
0:20
With Suno and Blender at hand, Claude Opus 5.5 worked for 3 hours 27 minutes and no keyframe was touched by hand. It first measured every pitch dive and silence in the track, carved the spectrogram into a magma canyon and made a pupil dilate on every kick.
676 saves on X
A 20-second animation whose landscape is the song's own spectrogram
Art, video & music
@kloss_xyz
733
42
676
0:20
Dmitriy Kovalenko, member of technical staff at OpenAI, took a clip posted with 'I'd love to see AI try and recreate this' and rebuilt it with Codex as an fframes project, with no video model involved.
2.1K saves on X
A viral motion clip recreated in Rust, SVG and shaders
Art, video & music
@neogoose_btw
2.2K
70
2.1K
3:01
A photocopy shop owner recruits a passing student to photograph an album cover, striking awkward poses with the confidence of a veteran producer. The three-minute AI short by Dan Pradana was made with Kling 4.0 Flash.
566 saves on X
A short film about an 83-year-old's musical debut in Jakarta
Art, video & music
@mxvdxn
1.2K
70
566
0:50
Opus 5.5 built it with procedural three.js WebGPU and TSL shaders only, with no Blender, assets or textures. The drawings are orthographic renders of the same model, showing elevation, pressure, trajectory and firing change in real time.
241 saves on X
An 1857 Napoleon cannon in procedural 3D, with live technical drawings
Art, video & music
@thebuggeddev
358
21
241
9:31
Left to run on its own for a day, Claude Code assembled the film from real archive footage and McLaren's own voice. The finished film, the workflow and every tool used are shared.
841 saves on X
A 10-minute Bruce McLaren documentary made by Claude Code in a day
Art, video & music
@maxescu
560
53
841
0:30
Sarah Parker, an AI creator, used Seedance 2.5 to stage a fictional 1980s production in which CRTs, floppy disks and primitive computers are presented as revolutionary AI. Everything plays it straight until the AI starts to glitch.
77 saves on X
A fake 1987 commercial for the world's first AI, until it glitches
Art, video & music
@SeharShinwari
306
7
77
1:08
Kevin Ngo's short film: a dad vibe codes a game and secretly makes the father's character weaker so the daughter always wins, until one day the daughter asks Claude why Dad always loses. Made with Sonnet 5.5 using real Claude Code sessions.
110 saves on X
Slow Me Down, a short film about a dad who always loses on purpose
Art, video & music
@kevin_t_ngo
366
23
110
0:33
A digital artist vibe-coded a web version of procedural graphics experiments first made in After Effects and Geometry Nodes, with presets built from the old shapes and plenty more to play with and break.
380 saves on X
Old After Effects experiments reborn as a procedural graphics toy
Art, video & music
@sashamartinsen
431
36
380
0:25
Made with Opus 5.5 working in Blender through an MCP server, Seedance 2.5 for video generation, and DaVinci Resolve Studio, which Opus 5.5 also operated.
3.4K saves on X
A retro anime sequence with Claude driving Blender and DaVinci Resolve
Art, video & music
@AIWarper
4.4K
307
3.4K
35:37
Given a few constraints and the API keys in the author's keychain, Opus 5.5 scripted it, narrated it with Gemini 3.8 Flash TTS, scored it with Lyrica 3, made the imagery and motion graphics, and wrote its own credits. It spent over $200.
2.2K saves on X
A 36-minute history-of-light cartoon that Opus wrote and produced
Art, video & music
@LLMJunky
2.2K
262
2.2K
0:32
Biba, a designer at Quaev Studios, is learning to make social media launch videos and is trying Claude Code as the tool. This 32-second spot for Lagos Run is one of the results.
113 saves on X
A designer's product launch video, made with Claude Code
Art, video & music
@UIbyBiba
712
60
113
3:20
Asked for a guitar riff, Opus 5.5 started a band, The Context Windows. It composed with GarageBand instruments by reverse engineering the file format, got the open SoulX-Singer model singing on MLX, and animated physically accurate fretting in Blender.
145 saves on X
An AI band's music video, made entirely on one MacBook in 24 hours
Art, video & music
@anshuc
273
30
145
3:09
Claude Fable 5.5 was asked for a video of all human progress, then 30 years into the future. It came back with the full animation and an original score, made without any footage or stock music.
5.2K saves on X
A 3-minute animated film of human progress from a single prompt
Art, video & music
@imjustnewatai
8.1K
784
5.2K
1:56
No models, textures or hand-made rigs: each animal is sculpted from SDFs, then meshed, rigged, skinned and furred at runtime and animated with IK. Made with Opus 5.5 in three.js; open source, with a live demo.
257 saves on X
24 animal species generated entirely from code, fur and rigs included
Art, video & music
@majidmanzarpour
312
17
257
0:51
Dhruval had Opus 5.5 build a Turkish marbling (ebru) engine from scratch, simulating ink drops, steel styluses and swirls at roughly 750 math operations per pixel, then had it write the soundtrack. Rendered in 4K.
248 saves on X
A Turkish marbling engine at 750 math ops per pixel
Art, video & music
@dhruvalgolakiya
475
29
248
0:29
Omah had Claude sync three camera feeds, cut the 40-minute episode straight from the transcript, and build all the graphics with Remotion, using few tokens and little prompting.
2.3K saves on X
Claude edited a 40-minute, three-camera podcast from the transcript
Art, video & music
@TassyOmah_
870
66
2.3K
4:00
After 10+ hours of tuning, Opus 5.5 writes the script, makes keyframes and clips through OpenRouter, narrates with Gemini TTS, adds ElevenLabs music, word-timed captions and a critic pass, then assembles the cut. The example explains neoclouds.
4.4K saves on X
A 4-minute explainer video made end to end in Claude Code
Art, video & music
@deedydas
2.1K
144
4.4K
0:42
Claude Sonnet 5.5 and GPT-6 Astra both got the one-line challenge "Build what you think you look like", and the video shows what each of them built.
3.9K saves on X
Self-portraits built by two rival models from the same prompt
Art, video & music
@siyabuilt
33K
390
3.9K
1 / 4
Asked by pietro, who is building Manufact, Codex turned SpaceX's four photos of Starship's first orbital flight into a dynamic macOS wallpaper. The file is posted for anyone to download.
3.3K saves on X
SpaceX's Starship liftoff photos turned into a dynamic Mac wallpaper
Art, video & music
@pietrozullo
3.6K
140
3.3K
0:05
A silent edit went into Codex and the agent designed original sound effects to fit it, with no DAW and no third-party samples. GPT-6 Astra/Sol did the work, and the prompt for building the sound pack is shared.
956 saves on X
Sound design for a muted video, made by a coding agent without a DAW
Art, video & music
@AmirMushich
667
39
956
0:33
Claude Code with Opus 5.5 worked for about 8 hours, inventing a fake calendar app along with its name, logo, UI, animations, soundtrack and sound effects. The repo is open source.
1.4K saves on X
A launch video where every frame and sound is code, from one prompt
Art, video & music
@LexnLin
1.4K
42
1.4K
0:47
Opus 5.5 turned a creative director's own Midjourney images into a single animation, with music from Suno. An earlier attempt at the same idea in After Effects crashed the project.
366 saves on X
More than 1,000 Midjourney images animated into one motion piece
Art, video & music
@ciguleva
730
18
366
1:05
Yunfan Ye, an AI researcher formerly at Google and Meta, had GPT-6 Astra model a house from its Zillow photos and cut a promo video, then reran it on an ordinary house with Claude Opus 5.5 in one shot. They call the result indistinguishable from a real walkthrough, and the code is open source.
340 saves on X
Zillow listing photos turned into a 3D house walkthrough video
Art, video & music
@realYunfanYe
323
20
340
0:29
Zero built a 2D product launch video entirely in code with Claude Opus 5.5, one continuous take where each scene morphs out of the last, synced to a 120 BPM track, with no external motion tools.
3.8K saves on X
A code-only Apple-keynote-style launch film made for $0
Art, video & music
@twoclipping
2.2K
160
3.8K
0:12
Built with Opus 5.5, the site's music is not a playlist: a brand-new track is composed in the browser every few minutes, set against pixel-art cities at night. Made to leave on while working.
1.4K saves on X
Endless lofi composed live in the browser over pixel-art cities
Art, video & music
@SafaElmali
1.6K
112
1.4K
1 / 2
The creator directed Claude Opus 5.5 in Claude Code, which planned 211 shots from their script, wrote a custom 2D animation engine, cast 11 ElevenLabs voices for 70 lines of Arabic dub and built the mixer, subtitles and render pipeline. 300+ generated images, 16 music cues.
64 saves on X
An 8½-minute Arabic-dubbed anime pilot produced in about 14 hours
Art, video & music
@sbalhatlani
73
8
64
0:14
Opus 5.5 made the 14-second video. The shared prompt template behind it has one shape morph without a cut from button to loader, music player, slider, toggle, tabs and a self-drawing chart, landing on every beat of a 120 BPM track.
21K saves on X
A UI motion design loop built entirely in code, no After Effects
Art, video & music
@twoclipping
13K
757
21K
0:30
Ole had Claude Opus 5.5 make a 30-second animated short of four seasons passing outside a cozy train window using only about 2,800 lines of code and no image or video model, then explain how it rendered and scored each frame.
193 saves on X
A 30-second animated film from 2,800 lines of code, no image model
Art, video & music
@itsolelehmann
274
19
193
0:17
Gregor Zunic gave Opus 5.5 with video-use thirteen takes of one sentence. It read every transcript, compared the top three frame by frame, picked the clean take, then cut, graded and captioned it, rewriting its own captions, in two minutes for 90 cents.
384 saves on X
An agent picked the best of 13 takes and edited the video
Art, video & music
@gregpr07
304
8
384
0:43
Asked to make the most impressive demo of itself, Claude Opus 5.5 worked for about 5 hours and 697M tokens. Every pixel and sound is generated in code, with no images, audio files, 3D models or libraries.
827 saves on X
A demoscene intro in one 280 KB HTML file, made with zero direction
Art, video & music
@JustinPerea
1.6K
87
827
3:10
Everything seen and heard in 'No Samples' runs on custom JavaScript that Opus 5.5 wrote, and Claude raps about how it all works.
869 saves on X
A rap single and music video made entirely in JavaScript
Art, video & music
@aj_dev_smith
2K
179
869
0:10
Victor M, head of product at Hugging Face, had Opus 5.5 write a self-contained Canvas 2D animation with no images or libraries: 128 by 96 pixels, legs driven by inverse kinematics and a 12-pose gallop cycle.
574 saves on X
A galloping pixel horse drawn procedurally in one HTML file
Art, video & music
@victormustar
1.1K
51
574
5:24
Davo made the short film 'The Last Call' with Seedance 2.5, ChatGPT and CapCut for roughly $38 in credits, and shared how the characters, thumbnail and shots came together.
654 saves on X
An AI short film made for about $38 in credits
Art, video & music
@DavoCreates
523
69
654
1:13
Scott asked Claude Opus 5.5 to turn a Tesla Cybertruck into a transformer, and it produced the full transformation as a three.js animation, shown in a 73-second clip.
72 saves on X
A Tesla Cybertruck that transforms into a robot, animated in three.js
Art, video & music
@scottstts
417
20
72
0:19
Built with Claude Opus 5.5 using three.js and TSL, the architecture moves through four stages: first lines, massing, detail and the finished home.
2K saves on X
A house that grows from sketch lines into a finished home in 3D
Art, video & music
@techartist_
2K
153
2K
0:28
Claude Opus 5.5 drew the whole thing in code. Everyone in town sends Claude their requests, but one girl sends a question instead: "What do you love?"
3K saves on X
An animated short about Claude, every frame drawn in JavaScript
Art, video & music
@kevin_t_ngo
6.2K
383
3K
1 / 4
Glitter Gal, an Adobe Firefly Ambassador who builds cute and whimsical worlds, shares the exact Midjourney recipe behind the images: three weighted style references, style weight 220 and stylize 350.
222 saves on X
Four Midjourney stills from a three-part style-reference blend
Art, video & music
@GlitterPixely
1K
205
222
1:40
Pradeep Kapoor asked GPT-6 Astra at max effort to paint a complex picture from scratch with nothing but code deciding where each stroke goes. No image generator, no reference and no picture underneath, just procedural painting stroke by stroke.
560 saves on X
A painting made of 68,000 brush strokes placed by code alone
Art, video & music
@pradeepXkapoor
974
40
560
3:16
Kelly Boesch, a digital and AI artist, wrote the song a while back and reworked the lyrics with Marshall Altman, who added production. The track was made in Suno, and the visuals try something different, with Midjourney images animated in Veo and Runway.
70 saves on X
An older song remade with Suno, Midjourney, Veo and Runway visuals
Art, video & music
@kellyeld
381
81
70
1 / 3
Product designer Sofiia Matsiutsia shares three experiments in a Risograph print style, generated with the GPT 2.5 Sunburst model inside ElevenLabs.
768 saves on X
Risograph-style prints generated with GPT 2.5 Sunburst
Art, video & music
@SMatsiutsia
1.3K
37
768
GIF
Curious whether a language model could draw like a pixel-predicting image model, Logan Bolton, an MS student at NYU Courant, gave Claude Opus and Fable the same prompts and had them output every pixel's RGB values into a .txt file. Opus came out looking much better, to their surprise.
156 saves on X
Pictures drawn by two LLMs emitting raw RGB values as text
Art, video & music
@septisum
557
15
156
19:37
Marco Magario, an AI filmmaker, wrote, directed, produced and scored the 20-minute piece alone and calls it their most ambitious AI short film.
877 saves on X
A 20-minute AI short film retells Plato's Myth of Er as dark comedy
Art, video & music
@MarcoMagarioAI
1.1K
157
877
1:27
Derrick Choi, who leads Codex for APAC at OpenAI, had Codex on GPT-6 Astra write the scripts that built and animated all 22 cars across 2,610 frames at 4K 30fps, and a Python-generated stereo soundtrack of engine revs, gear changes, passing cars and tunnel reverb.
167 saves on X
An 87-second Monaco Grand Prix final lap rendered from code in Blender
Art, video & music
@derrickcchoi
405
15
167
0:46
AI filmmaker Víctor Sánchez of VoxelPlot generated the full fantasy duel with Wan 3.0, Seedance 2.5 and Hailuo H3, passing shots and references between them with Vid2Vid and Omni Reference. Wan handled timing and layouts; Seedance kept the characters on-model.
305 saves on X
A mage-versus-swordswoman anime fight cut from three video models
Art, video & music
@voxelplot
434
34
305
0:27
Kevin Ngo, who does creative coding with AI, had Claude Opus 5 write the JavaScript that draws every frame of a 27-second animated short about a fruit fly's life.
3.9K saves on X
The life of a fruit fly, animated frame by frame in JavaScript
Art, video & music
@kevin_t_ngo
6.8K
462
3.9K
0:30
Arojinle, a wildlife and conservation researcher, put Àjànàkú ò Ṣièrè on YouTube, a Yoruba-language series made with AI. The opening story turns on forest elephants that live near Eringbó village and had never wandered into it.
142 saves on X
A Yoruba AI series about elephants near a village
Art, video & music
@arojinle1
1K
266
142
Historian Tom Rowsell asked ChatGPT for a portrait of Benjamin Rowsell, a 3rd great-grandfather who was a City stockbroker in the 1860s, working from a drawing by Benjamin's six-year-old son and photos of two sons as adults. Rowsell admits nobody who knew the man would recognise it.
11 saves on X
A portrait of an ancestor with no photos, built from a child's drawing
Art, video & music
@Tom_Rowsell
127
2
11
Alessandro Riolo's sailor grandfather kept the family's rooftop tanica clean the old Sicilian way: a live eel that ate algae and larvae. With the memory 80 years old and the practice gone, they had ChatGPT draw the scene their mother still laughs about.
136 saves on X
A grandfather's rooftop eel, drawn from an 80-year-old family story
Art, video & music
@aledeniz
575
57
136
1 / 4
Alban Glazman, a freelancer, ran Homer's Odyssey through Midjourney in its high-definition mode and posted four scenes from the epic.
1.3K saves on X
Scenes from the Odyssey drawn in Midjourney
Art, video & music
@alban_gz
5.5K
451
1.3K
0:18
Adil Alimzhanov asked GPT-6 Astra to colour grade raw gym footage inside Premiere Pro. The 18-second result plays like the opening of a boxing film.
543 saves on X
Gym footage graded in Premiere Pro to feel like a boxing film
Art, video & music
@adilinthewild
764
47
543
1:02
Szymon Rączka kept noticing the time 10:09 on screens during the Apple keynote, so they asked GPT-6 Astra to compile every instance. It came back with 50 shots all reading 10:09, ending on the iPhone Duo showing 9:41.
1.2K saves on X
A 50-shot supercut of every 10:09 in the Apple keynote, cut by GPT-6
Art, video & music
@screenfluent
2.9K
53
1.2K
AI filmmaker M. Asif prompted ChatGPT Image 2.5 for a studio-style production board: hero portrait, five-view turnaround, six expressions, pose studies, costume callouts, material studies and a colour palette, all locked to one character. The full prompt is in the post.
3.5K saves on X
A full character model sheet from one reference image
Art, video & music
@meAsifAi
1.8K
233
3.5K
3:34
Jeffrey Emanuel's closed-source system pairs mtdt, a Rust CLI that imports, exports and manipulates MIDI, MusicXML, PDF and audio, with over 300 music-theory skills. Driven by GPT-6 Astra it composed the fugue and produced the print-ready score with no other software.
334 saves on X
A music-theory system composed a Bach-style fugue and its score
Art, video & music
@doodlestein
505
40
334
0:45
Spenser Dickerson, a product marketing manager, spent about 10 hours directing GPT-6 Astra in Blender on a MacBook Pro. Asked for a hoop and flooring built to spec, it sourced manufacturer drawings and blocked everything out at the right scale. Lots of art direction, not one prompt.
283 saves on X
A vintage gym built in Blender over 21 rounds of revisions
Art, video & music
@SpenserFX
423
35
283
0:38
Blendi, previously an engineer at fal and a co-founder of Remade, made a non-stop AI shopping channel. Viewers invent the products and the host, Graham, has to sell them. It runs on H3 Max turbo at 1080p in real time.
340 saves on X
An infinite shopping channel that sells products you invent
Art, video & music
@BlendiByl
521
37
340
0:54
Ramon, a pianist and developer, asked GPT-6 Astra to compose a piano piece in a late-Romantic style and was taken aback by the harmonies it came up with.
801 saves on X
A late-Romantic piano piece composed on request
Art, video & music
@rvm0n_
1.5K
75
801
0:43
Timour Kosters, co-founder of Edge City, gave GPT-6 Astra one instruction, "use Ableton Live and make an absolute banger", and let it operate the DAW on its own. The 43-second clip is what came out.
3.5K saves on X
One prompt sent GPT-6 Astra into Ableton Live to make a track
Art, video & music
@timourxyz
9.1K
815
3.5K
3:49
High-school student Leon Lin asked GPT-6 Astra to compose a piano piece and build a 3D pianist to perform it. The hands track the actual notes as they sound, the whole thing took about two hours, and it runs in the browser.
374 saves on X
A piano melody played by a 3D pianist whose fingers follow the notes
Art, video & music
@LexnLin
1.1K
83
374
2:00
Derya Unutmaz, a professor and immunologist, asked GPT-6 Astra for a fully playable browser piano and then had it build in Bach's Brandenburg Concertos. The whole thing took about 11 minutes; all six concertos can be played straight on the keys at the link.
676 saves on X
A playable virtual piano with all six Brandenburg Concertos built in
Art, video & music
@DeryaTR_
1.5K
190
676
MarsForTech gave GPT-6 Astra Max one line: generate an SVG of a PlayStation 4 controller as nicely as you can. 11 minutes 54 seconds later it had drawn the whole thing from code, with no image model involved.
852 saves on X
A PS4 controller drawn as pure SVG code in 12 minutes, no image used
Art, video & music
@MarsForTech
5.2K
78
852
0:34
Tom Krcha, founder of Pendev, handed GPT-6 Astra an old drawing of a steam train and asked for a Blender reconstruction. A few minutes later it had modelled 3,295 fully editable, detailed objects, and you can tell it how simple or detailed to go.
3.5K saves on X
Old steam-train drawing rebuilt in Blender as 3,295 editable objects
Art, video & music
@tomkrcha
6.9K
498
3.5K
1 / 2
Mahnoor Fatima, a generative AI creator, gave ChatGPT a waterfront photo and a long prompt asking for a two-panel image: the untouched original above, and the same scene below as an ink-and-watercolour print on ivory paper with typewriter captions.
234 saves on X
A photo redrawn as a vintage travel-journal print, original on top
Art, video & music
@MahnoorAi12
239
26
234
0:20
Tom Krcha, founder of Pendev, gave GPT-6 Astra an image of a house and asked for it in 3D with every detail, down to toys, appliances and furniture. The result is a complete Blender reconstruction with tweakable geometry that also runs at 60fps as a locally rendered game.
3.8K saves on X
One photo of a house became a full, editable 3D model in Blender
Art, video & music
@tomkrcha
6.5K
564
3.8K
0:09
Pietro Schirano, CEO of MagicPath and formerly at Anthropic, had early access to GPT-6 and started sharing tests. In the first, the model turned one image into a 3D model with an animation, all written in code rather than rendered by an image or video model.
6.7K saves on X
A 3D model and animation written in code from a single image
Art, video & music
@skirano
10K
582
6.7K
0:29
Rahul Nanda, a filmmaker and former Times of India journalist with 19 years in multimedia, made this short Bollywood romance entirely with Seedance 2.5.
797 saves on X
A messy Bollywood love story in 29 seconds
Art, video & music
@rahulnanda86
1.2K
68
797
0:19
Product designer Yan Liu, already missing summer in September, used Claude to create a short piece of sailing across a lake, green water in sunlight and waves on a beach, in crisp colours.
706 saves on X
Lake, sunlight and waves: a summer animation made with Claude
Art, video & music
@yanliudesign
782
35
706
1 / 4
OscarAI, an artist who works with generative AI, shares four pictures in a charming vintage illustration style they developed in Midjourney the day before.
114 saves on X
A new vintage-feel illustration style made in Midjourney
Art, video & music
@Artedeingenio
183
16
114
1 / 2
Taaruk shared two GPT Image 2 prompts and their results: a high-key editorial photograph of an empty city train with one orange seat, then the same scene redrawn as an ink-and-watercolour page on aged ivory paper, titled 'City Ride'.
1.3K saves on X
One train carriage rendered as a photo and as a travel-zine sketch
Art, video & music
@Taaruk_
1.2K
143
1.3K
4:49
Mark Puchinsky, a storyteller with credits on Law & Order: SVU and The Blacklist, adapted Book 24 of the Odyssey, where father and son meet again after twenty years, for the Grok Imagine contest. Nostos is their first short with any generative tool.
27 saves on X
A short film of Odysseus and Laertes reunited after 20 years
Art, video & music
@MarkPuchinsky
269
13
27
1 / 3
HeisenKeyX used ChatGPT to repair an old, damaged childhood picture of their mum and big mum. The before and after are in the post, and both were on cloud nine when shown the result.
22 saves on X
A damaged childhood photo of mum and big mum, restored
Art, video & music
@KNandd
184
9
22
6:13
Julien Prevost, a French film director with 15 years on real sets, made Incubus in two weeks for the Higgsfield festival: a story they had carried for eleven years and call unfilmable until now.
1K saves on X
A six-minute AI short about a man who can live inside other people
Art, video & music
@JulienPrevostFR
1.6K
235
1K
1 / 4
Saul Goodman's GPT Image 2 prompt asks for a compact multi-colour stamp of each place's most recognisable shapes, with dry-ink gaps, slight misregistration and typewriter captions on off-white paper. The post shares four finished posters and the full prompt.
1.1K saves on X
Travel posters that look like hand-carved rubber stamps on aged paper
Art, video & music
@Goodmanprotocol
1.1K
131
1.1K
1 / 2
Laraib Fatima turns a travel photo into a split-screen poster with GPT Image 2 in ChatGPT: the original untouched on the left, a matte background sampled from it on the right with an enamel-pin badge of the scene and the place name in serif. First of a twelve-part series, Collected Journeys.
228 saves on X
Luxury travel posters made from the author's own trip photos
Art, video & music
@AiwithLariab
292
33
228
1 / 2
Mahnoor Fatima closes a 12-part series with GPT Image 2 in ChatGPT: the top half of each page is a cinematic travel photo, the bottom turns the same pose into loose black-and-white ink on cream paper with diary notes. The prompt is in the post.
295 saves on X
Departure Diary: a travel photo meets an ink sketch on one journal page
Art, video & music
@MahnoorAi12
296
32
295
1 / 4
Palak, who takes on design work, shared four images generated with ChatGPT as the visual material for a recent design project.
326 saves on X
Imagery for a recent design project generated in ChatGPT
Art, video & music
@Palakonweb
568
19
326
4:36
Kelly Boesch, a digital and AI artist, wrote the lyrics and generated the song in Suno, then Marshall Altman added production and guitar. This is the second video made for the track, which is out on streaming services.
571 saves on X
'Not Made For The Cage', a Suno song finished with live guitar
Art, video & music
@kellyeld
1.6K
316
571
4:15
biglootpig, co-founder of a generative media tooling startup, produced the four-minute episode with a custom agent harness they call beancat, Seedance 2.5 and the Atlas Cloud API. The whole series is on a YouTube playlist.
3.2K saves on X
Episode 5 of Atlas, an agent-made anime series in 80s and 90s style
Art, video & music
@beeeeeeeegyoshi
5.5K
311
3.2K
0:46
Two spies fight until the camera pulls back to reveal a video game, and the player's mom walks in. bot. creates made the 90s OVA-style short from Midjourney reference images animated in Seedance 2.5.
1.5K saves on X
Final Boss: Mom, a retro-anime spy fight that ends with a chancla
Art, video & music
@botcreates
3.1K
225
1.5K
1 / 3
simeon-sanai's GPT Image 2 prompt asks for a country's landmarks, food, transport and street life to flow into one hand-drawn sketch: bold black ink, cross-hatching, off-white paper, two or three accent colours and handwritten labels. Swap in any country name.
762 saves on X
Country travel posters drawn as one continuous ink line
Art, video & music
@Naiknelofar788
829
115
762
0:29
Burak Tuyan wrote the script by hand, had ChatGPT 5.6 Sol Max turn it into a 14,000-character Seedance 2.5 prompt and ran 20+ tests at 480p before the final 720p renders. GPT Image 2 made the trial-screen reference, Seed Audio 1.0 the voices, Topaz the upscale.
525 saves on X
A 30-second WinRAR spec ad about the 40-day trial that never ended
Art, video & music
@buraktuyan
2K
222
525
0:20
Bryl Lim, a founder and principal AI engineer, used Codex with image generation to produce more than a thousand SVGs for their workout app, scrolling past in the demo.
3.6K saves on X
Over 1,000 SVG illustrations generated for a workout app
Art, video & music
@bryllim_
3K
112
3.6K
0:33
Ann Nguyen asked Claude Opus 5 to explain a tweet-sized p5.js animation by another artist. It worked out that every point was placed by math, then wrote its own piece in the same style, with squids.
1.5K saves on X
Squids in p5.js, after Claude broke down how another sketch worked
Art, video & music
@ann_nnng
1.9K
139
1.5K
No matches.
Show everything
Show more
