Meshy v6 — Image or Text to 3D, Your Way
3D Models from an Image or a Prompt, in Multiple Formats
Meshy v6 turns either a single image or a plain-text description into a complete 3D model. Upload a product shot or character concept and the image-to-3D pipeline reconstructs the mesh and textures it — or skip the image, describe what you want ('a stylized low-poly red fox, game-ready'), and the text-to-3D pipeline builds the model from language alone. Either way you can pick a standard or low-poly mesh, choose triangle or quad topology, set a target polygon count, and turn on PBR materials so the surface reacts correctly to lighting. Export the result as GLB, FBX, OBJ, USDZ, or STL, and preview it interactively right on this page before you download.
The two input modes cover different starting points. Image-to-3D is the right call when you already have a reference — a photo, a render, or concept art — and want the model to follow it; a single clear image with the subject separated from the background gives the cleanest result. Text-to-3D is for when you're starting from an idea rather than a picture: write a concrete prompt naming the object, its style, and the details that matter, and the model composes geometry and textures from the description.
Meshy v6 gives you direct control over the mesh. Choose the standard style for general-purpose assets or low-poly for stylized, lightweight game models. Pick triangle topology for engine-ready meshes or quad topology for cleaner edge flow in DCC tools, set a target polygon count to match your budget, and use symmetry control to keep characters and props balanced. Add a texture prompt to steer the surface look, and enable PBR materials when you want the model to respond to lighting in modern engines and glTF viewers.
The workflow stays simple. Pick a mode, provide your image or prompt, adjust the mesh and texture options if you want, then generate and inspect the model in the built-in interactive viewer — orbit, zoom, and check the silhouette from every angle. When it looks right, download it in the format your pipeline needs. Every generation is also saved to your dashboard gallery so you can return to it later.
How to Generate a 3D Model with Meshy v6
Pick a Mode & Provide Input
Choose Image to 3D and upload a clear reference image, or switch to Text to 3D and describe the model you want — name the object, its style, and the key details.
Set Mesh & Texture Options
Pick a standard or low-poly style, choose triangle or quad topology, set your target polygon count, and toggle PBR materials — or add a texture prompt to steer the surface look.
Generate & Download
Click Generate, then orbit and zoom the result in the interactive viewer. Download it as GLB, FBX, OBJ, USDZ, or STL — every model is also saved to your dashboard gallery.
Meshy v6 Technical Specifications
| Provider | Meshy |
| Platform | Cloud API (hosted) |
| Input Modes | Image-to-3D and text-to-3D |
| Input | Single image, or a text prompt (up to 600 characters) |
| Mesh Style | Standard or low-poly |
| Topology | Triangle or quad |
| Polygon Control | Adjustable target polygon count (default 30,000) |
| PBR Materials | Optional |
| Output Formats | GLB, FBX, OBJ, USDZ, STL |
| Rigging & Animation | Optional — auto-rigging with preset animations |
| Preview | Interactive in-browser 3D viewer |
| Commercial Use | Supported |
Why Choose Meshy v6
Image or Text to 3D
Start from a reference photo or concept art with image-to-3D, or from a plain-language prompt with text-to-3D — one model, two ways in, the same controllable mesh output.
Control the Mesh
Choose standard or low-poly, triangle or quad topology, and a target polygon count, with symmetry control and optional PBR materials — tune the model to fit your engine or DCC pipeline.
Export Anywhere
Download GLB, FBX, OBJ, USDZ, or STL and preview interactively before you commit — drop the model straight into game engines, AR runtimes, DCC tools, or a 3D printer.
Meshy v6 vs Other Image-to-3D Models
| Feature | Meshy v6 | Hunyuan 3D V3.1 | Trellis 2 |
|---|---|---|---|
| Input | Image or text prompt | Image or text prompt | Single image |
| Mesh Style | Standard or low-poly | Standard | Clean topology |
| Topology | Triangle or quad | Triangle | Triangle |
| PBR Materials | Optional | Optional | Texture only |
| Export | GLB, FBX, OBJ, USDZ, STL | GLB + OBJ | GLB |
| Best For | Game & AR assets | High-fidelity assets | Clean-topology models |
What Can You Create with Meshy v6
Game Assets
Turn concept art, reference photos, or a written brief into props, characters, and environment pieces — pick low-poly and a polygon budget tuned for real-time engines.
AR & VR
Produce lightweight GLB and USDZ assets that drop straight into AR experiences and VR scenes, viewable in any glTF- or USDZ-compatible runtime.
Product Visualization
Convert a single product photo into an interactive 3D model for e-commerce and configurators — with optional PBR materials for accurate surfaces.
3D Printing
Generate a mesh from an image or a prompt and export STL or OBJ for slicing, prototyping, and physical models of objects, miniatures, and collectibles.
Idea or Concept to 3D
Have a 2D concept? Use image-to-3D. Have only an idea? Describe it with text-to-3D. Either way, get a rotatable 3D version in minutes — ideal for pitching, previs, and iteration.
Animation & Film
Build set-dressing props and background assets fast, then refine in your DCC tool of choice with quad topology and clean FBX or OBJ exports.
Related AI Models
Frequently Asked Questions About Meshy v6
What input does Meshy v6 accept — images or text?
Both. Use the Image to 3D / Text to 3D switch at the top of the generator. Image-to-3D reconstructs a model from a reference picture, while text-to-3D builds one from a written description of up to 600 characters. Both modes produce a textured 3D model you can export in several formats.
What's the difference between standard and low-poly?
Standard produces a general-purpose mesh suitable for most uses, while low-poly produces a lighter, stylized mesh with fewer polygons — a good fit for real-time games and stylized art. You can also set a target polygon count to fine-tune the budget either way.
Should I use triangle or quad topology?
Triangle topology is engine-ready and works everywhere out of the box. Quad topology gives cleaner edge flow, which is preferable when you plan to edit, subdivide, or animate the model in a DCC tool like Blender or Maya.
What are PBR materials and should I enable them?
PBR (physically based rendering) materials describe how a surface reacts to light. Enable PBR when you want realistic results in engines like Unreal, Unity, Blender, or three.js. Leave it off if you only need quick geometry or plan to texture the model yourself. You can also add a texture prompt to steer the surface look.
Which formats can I export?
Meshy v6 outputs GLB for the web, game engines, and AR, plus FBX and OBJ for DCC tools, USDZ for Apple AR, and STL for 3D printing. The model is also previewed interactively in your browser before you download.
How is this different from Hunyuan 3D V3.1 and Trellis 2?
All three turn inputs into 3D. Meshy v6 stands out for direct mesh control — standard or low-poly, triangle or quad topology, target polygon count — and a wide set of export formats. Hunyuan 3D V3.1 emphasizes high-fidelity geometry with GLB and OBJ output, while Trellis 2 is image-only and focuses on clean topology. Try them on the same subject and keep the result that fits your pipeline.
Can Meshy v6 rig and animate a character?
Yes — enable auto-rigging to add a skeleton (set the character height for correct proportions), then optionally apply a preset animation by action id. When rigging or animation is on, the previewed and downloaded model is the rigged or animated GLB. Rigging and animation are billed as add-ons on top of the base generation.