Meshy web app documentation
Create your first 3D model in the Meshy web app from an image or a text prompt, then download it for your project.
This guide is for new Meshy users. To build Meshy into your own code instead, see the API Quickstart. To create through a chat, see 3D Agent.
Create your first 3D model
To create your first model, follow these steps:
-
Go to meshy.ai and sign up with Google or email. New accounts get free credits.
-
In the left sidebar, click Model.
-
At the top of the panel, choose how to generate:
-
Image to 3D: upload a clear photo or concept art of your subject. This is the best place to start.
-
Batch Images to 3D: upload up to 10 images and turn each one into a model at once. Batch Images to 3D requires a paid plan.
-
Text to 3D: describe your object in detail, for example "a medieval wooden treasure chest with iron bands and a rusty padlock".
-
-
Select High Detail or Smart Topology.
-
In the model list, select an AI model:
-
High Detail offers Meshy 7.1 - Flagship, Meshy 6, and Meshy 6 Lite. Text to 3D offers Meshy 6 and Meshy 6 Lite.
-
Smart Topology offers Meshy T2 and Meshy T1. Text to 3D uses Meshy T1.
-
-
Optional: Set the options you need, such as Multi-view, Resolution, Texture, or Image Enhancement. The panel shows only the options the selected model supports.
-
Click Generate, or Batch Generate for Batch Images to 3D. The panel shows the time and credit cost before you generate.
-
In the viewer, rotate, zoom, and check the mesh quality, texture detail, and overall accuracy.
-
Optional: Improve the model:
-
To texture it in a different style or add PBR maps, click Texture or Retexture in the toolbar.
-
To change the polygon count, click Remesh in the toolbar. Choose a count for your target platform, such as a game engine or a 3D printer.
-
To animate a humanoid or four-legged model, click Animate in the left sidebar, and then click Rig to auto-rig it.
-
-
In the toolbar, click Download.
-
In Format, select a format, and then click Download. Use GLB for web and AR, FBX for game engines, STL for 3D printing, or OBJ for DCC tools.
Settings and tradeoffs
The following table shows what the main generation settings trade off.
| Setting | Options | Tradeoff |
|---|---|---|
| AI model | Meshy 7.1 - Flagship, Meshy 6, or Meshy 6 Lite | Meshy 7.1 gives the closest match to the input image. Meshy 6 suits models where a stable structure matters more than fine detail. Meshy 6 Lite is faster and costs 10 credits. |
| Model type | High Detail or Smart Topology | High Detail gives a detailed mesh. Smart Topology gives fewer faces and cleaner topology, and Meshy T2 generates in about 2 seconds. |
| Image Enhancement | On or off | On: Meshy redraws and cleans up the input image. Off: Meshy uses the image as it is. |
Examples
A clean input gives a clean model. A clear front-facing photo of a ceramic mug, generated with Meshy 7.1, High Detail, and Texture on, gives a clean mesh with an accurate handle, a smooth surface, and textures.
A poor input gives a poor model. A blurry, low-contrast photo with a complex background produces distorted geometry, because the model picks up the background. To fix it, keep Image Enhancement on, or use a cleaner input image.
Troubleshooting
The following table lists common problems, their causes, and how to fix them.
| Problem | Cause | Fix |
|---|---|---|
| Generation stays at 0% | Meshy's servers are busy | Wait 1 to 2 minutes, and then try again. Check the Meshy status page. |
| The model has holes or missing parts | The input image shows only part of the object | With Meshy 7.1, turn on Multi-view and add images from other angles. |
| Textures look blurry | The input image has a low resolution | Turn on Image Enhancement, or use a higher-resolution image. |
| The downloaded file is empty | The generation didn't finish | Wait until the generation finishes, and then download the model again. |
FAQ
- How do I create my first 3D model in Meshy?
- Sign up at meshy.ai, click Model in the left sidebar, and open Image to 3D or Text to 3D. Upload an image or enter a prompt, select a model, and click Generate.
- How long does it take to generate a 3D model?
- About a minute in most cases. Meshy 7.1 and Meshy 6 take about a minute, and Meshy T2 takes about 2 seconds. Ultra 4K resolution takes longer.
- Should I start with Text to 3D or Image to 3D?
- Use Image to 3D if you have a reference photo, because it gives the most accurate result. Use Text to 3D when you only have an idea in mind.
- Do I need a credit card to start?
- No. You get free credits as soon as you sign up with Google or email.
- Where are my generated models saved?
- Meshy saves every result automatically. Find it in the asset list next to the viewer, or click Assets in the left sidebar. From there you can preview, post-process, and download it.
Next steps
-
To learn every option for image input, see Image to 3D.
-
To write prompts for text input, see Text to 3D.
-
To create through a chat, see 3D Agent.