Our 3D to Image tool lets you quickly generate high-quality images based on the 3D scenes you build. With AI, you can preserve layout and composition while creating more expressive visuals, which can then be extended into dynamic video content.
1. Accessing the 3D Editor
You can enter through:
Click 3D to Image in the Workspace
Or go directly to the tool page: www.meshy.ai/scene
2. Create or Open a Scene
In the scene management panel, you can:
Create a new scene: Start from a blank scene or a preset template
Open a saved scene: Resume editing previous work
Rename/Delete scenes: For easier management
3. Build Your 3D Scene
In the editor, you can:
Add models: Select from the asset library
Adjust models: Rotate, scale, and move with Gizmo controls
Adjust the camera:
Left-click drag = Orbit
Right-click drag = Pan
Scroll = Step zoom
Middle-click drag = Smooth zoom
Adjust the environment: Change background, lighting, or ground color
This step defines the spatial layout and composition of your final image.
4. Add Prompts and References
In the right-side generation panel:
3D Reference (auto): A snapshot of your 3D scene is used as a hard constraint for layout and perspective
Text Prompt (required): Describe style and details
(e.g. Medieval town square with a central cobblestone path leading to a large twin-towered castle gate, flanked by a vibrant fabric and food market on the left and a blacksmith forge on the right. Barrels, crates, and two wooden carts—one with hay, one with supplies—are scattered in the foreground. Balanced wide-angle front view, buildings and towers symmetrically positioned in the background.
Ultra-detailed textures, cinematic lighting, high dynamic range, enhanced material realism, slightly foggy medieval atmosphere.)
Image Prompt (optional): Upload extra reference images for characters, backgrounds, etc. (GPT-Image only)
💡 Tip: You can refer to specific images in your prompt, e.g. Use image #2 as background. (The default 3D Reference is image #1).
Input Fidelity：
Expressive: Produces more detailed and visually expressive images, but may deviate from the scene reference.
Consistent: Follows the scene reference more strictly, ensuring higher consistency.
5. Generate Images
Click Generate Image:
GPT Model: Higher quality, more expressive, but may deviate from references
Flux Model: Faster, more consistent with 3D depth, but requires precise prompts
After generation, results appear in the image list where you can:
Preview in full size
Reuse the prompt
Delete or save results
6. Generate Videos
You can enter through:
From any generated image, click Send to Video:
Click Image to video in the Workspace
Or go directly to the tool page:https://app-staging.meshy.ai/video
Choose duration (5s / 10s) and aspect ratio 1:1, 16:9, 9:16(text to video only)
The image will be used as the starting frame for video generation
Preview, download, or reuse prompts in the results
7. Saving & Revisiting
The system automatically saves your scene progress
You can always return to scene management to continue editing or generate from past scenes