This stylized grocery store was created with Meshy! ๐คฏ
The secret? Build it piece by piece!
๐ 2D concepts: Use Nano Banana in Meshy to turn your sketch into a stylized 2D concept.
๐ช 3D parts: Convert each piece into a high-quality 3D model with Meshy Image to 3D.
๐ Assemble: Combine everything in Unreal Engine for total creative control.
๐จโ๐จ Special thanks to 3D artist @BjLin10 for inspiration!
God and His Servant, black-and-white backlit silhouettes. An antlered god and his kneeling servant, a study in unequal gazes. made with GPT Image 2.5 #AIart#silhouette
Museum Encounters, silkscreen prints. A curator and a thief, rival collectors, a professor and his student, a restorer and the statue that woke up. made with GPT Image 2.5 #AIart#silkscreen
saw @chetanankola's cozy creek and had to try my own
a little three.js creek running live in the browser. flow, depth, leaves and murkiness are all sliders, and the water sound shifts from trickle to rapids with the current
built it over ~12 versions with GPT-6.1. i'm a 3D artist, so i did the looking, it did the code ๐
how it got there:
v3 glints shattered into noise, looked like a lake
v7 sun as a disk with real size, glints finally behaved
v8 bed stopped looking doubled, foam stopped looking like a net
v12 leaves ride the current around rocks
my job: saying "this looks wrong" and why
The full breakdown of this knight is up.
One image > Meshy 7 Ultra > a separate head model for a sharper face > Blender by hand: lash cards on the lid line, 79,004 hair curves, freckles where the sun puts them.
7 Meshy tasks, 150 credits.
This knight started from one image.
Meshy for the concept and the 3D model, a separate head model for a sharper face, then Blender by hand: freckles, lashes, 79,004 hair curves.
7 Meshy tasks, 150 credits.
One concept image. Can you get an editable scene in Blender? ๐งฑ
Here's the workflow:
Decompose โ GPT-6 plans the parts
Generate โ Meshy API builds 12 textured assets
Refine โ GPT-6 + Blender adjust scale and placement
Present โ lighting, 30fps turntable
I got to generate models for the Meshy 7 tech report, so I was one of the first people to use it.
Honestly, the output went straight into Blender and already looked good. All I did was light it and dial in the render.
That's a lot of time artists get back. Go try it.
๐ข Meshy 7 is Now Live: Pushing the Frontier of 3D Alignment
Alignment is the gap between your input image and the generated 3D model. Meshy 7 narrows it.
Bring your hardest reference and see how close it lands.
Introducing Meshy 3D Agent ๐ The world's first AI agent for 3D creation, now in Beta.
Just chat โ the agent will:
๐ง Brainstorm with you, pitching directions first
๐จ Generate visuals in batches, refined through chat
๐ง Turn your favorite into 3D โ print, download, anywhere
๐ Answer your 3D questions in-line
Create 3D in conversation with Meshy Now โ https://t.co/3g1voTIUhY
GPT Image 2 - Character Reference Sheet Prompt
You can create a character reference sheet in any style you want for use in your Seedance 2.0 videos. All you need to do is either provide a reference image or describe your character. You can also input random values to generate completely random characters.
This reference sheet provides a clear, structured, and highly readable visual breakdown of the character, helping the model maintain strong identity consistency across different outputs. The layout ensures the model focuses on essential visual information rather than stylistic noise, while the inclusion of expressions, head angles, and performance cues improves the modelโs ability to generate believable acting and varied poses without losing character integrity.
All clips in the demonstration video were generated using the reference sheet image and the same Seedance 2.0 prompt. You can find the prompts in the replies. ๐
GPT Image 2 and Seedance 2.0 are both available on @mitte_ai