# How to use the AI model generation feature?
The new AI model generation function includes two core modules: **Text to 3D** and **Image to 3D**. It aims to use generative AI technology to assist designers in quickly transforming inspiration into 3D assets and release creativity.
### Text to 3D
Text generation models enable content creators to convert natural-language descriptions into 3D assets quickly. Simply enter text to generate a model that meets your needs, greatly improving the efficiency of asset acquisition.
#### How to use
1. **Input Description:** Describe the object feature you want to generate in detail in the input box.
2. **Click Generate:** After confirming that the description is correct, click the "Generate" button to start the task.
3. **Use a Fixed Seed:** This function supports a fixed random seed. When you are satisfied with a build result and want to fine-tune or need to reproduce a specific effect, keep the same cue and random seed to get consistent results.
### Image to 3D
The newly launched image generation model function can intelligently analyze 2D image data, infer its spatial structure, and reconstruct it into a 3D model with rich detail in a short time.
#### 1. Upload the main-view image
The main picture is the core reference for AI to understand the structure of objects.
* Operation: Drag or click to upload a main-view for generation.
* Supported formats:`.jpg`,`.jpeg`, and`.png`
* File size: up to 10 MB
Upload images
💡 **Upload Guides:**
* The quality and clarity of the reference image has a significant impact on the generated 3D model.
* High-resolution images are recommended (preferably above 1040 x 1040 pixels).
* Images with simple backgrounds, sufficient lighting, and complete and clear display of the object are recommended.
#### 2. Multi-angle Views (optional)
For a more complete model structure, you can enable the multi-view option.
* Operation: Upload up to three pictures of the same object from different angles (such as side, back, top view, etc.).
* Principle: AI will comprehensively analyze all picture information and perform three-dimensional reconstruction from multiple angles. The inference of the generated model in the invisible area will be more accurate, and the details will be richer.
* Supported formats:`.jpg`,`.jpeg`,`.png`
Multi-angle Views
::: info Upload Guides
- All pictures should be of different angles of the same object for best results.
- Please avoid uploading pictures including multiple objects or mixed views to prevent AI confusion.
- Picture order does not affect generation results.
:::
#### 3. Generation Settings
* **Model Name:** After entering a name for the 3D model to be generated, click the Generate button to start a generation task.
* **Use a Fixed Seed:** Using the same prompt with a fixed seed produces consistent results, making it easy to reproduce specific models.
* **Start generating:** After the settings are completed, click the "Generate" button to create a task. **Precautions**
* **Testing:** This feature is currently in **Beta**. After clicking the function entrance, users can receive 10 free experience opportunities. After the points are consumed, they can earn more experience points by applying for a test qualification.
* **Credits cost:** Each generation task will consume a certain amount of credits.
