3D Agent
Chat with Meshy 3D Agent to brainstorm ideas, generate visual concepts in batches, and turn the concepts you pick into downloadable 3D models.
Start with a photo, a sketch, a description, a story, or a creative direction. Unlike a tool that makes one model from one prompt, 3D Agent keeps the whole creative process in one conversation. It suits makers with a custom idea, such as a pet photo or a child's drawing, and indie game developers who need a set of assets in one style.
For a precise model from a clear reference image, use Image to 3D. For a single model from a well-defined prompt, use Text to 3D. To generate models from your own code, see the Meshy API documentation.
Create a model with 3D Agent
To create a model in a conversation, follow these steps:
-
In the left sidebar, click Agent.
-
In the message box, describe what you want to create. You can start with any of these:
-
A text description: "I want to create a set of fantasy potion bottles for my RPG game"
-
A photo: upload a pet photo, a product image, or a sketch
-
A story or scenario: "My daughter drew a dragon, and I want to turn it into a 3D printed figurine"
-
A creative direction: "Cyberpunk props with neon accents and weathered metal"
-
-
Brainstorm with the agent. It suggests creative directions and asks questions to help you refine your concept. You can compare artistic approaches, ask practical questions about 3D creation or printing, and refine your direction over several messages.
-
Ask for visual concepts. The agent generates several concepts in one session. For example:
-
Ask for a number of concepts: "Show me 6 variations"
-
Keep a style: "Keep the same art direction as the first batch"
-
Combine concepts: "I like the shape from concept 2 and the colors from concept 5"
-
-
Select the concepts you like, and ask the agent to convert them to 3D models. By default, the agent asks you to approve an action before it spends credits.
-
Download the model in the format your game, 3D printer, store, or other workflow needs.
Key capabilities
3D Agent combines these capabilities in one conversation:
-
Brainstorming for 3D creation: the agent suggests creative directions before you generate anything. This helps most when you know the kind of object you need, such as game props, but not the specific designs.
-
Batch visual concepts: instead of one result per prompt, you explore several concepts in one conversation. The agent can keep the style consistent across batches, which matters for asset sets that belong to the same world.
-
Chat-to-3D creation: the agent turns the concepts you select into downloadable 3D models for creative, printing, or production work.
-
Built-in 3D knowledge: ask questions about 3D creation and 3D printing in the chat. The agent can advise on topics such as wall thickness for printing, polygon budgets for different platforms, or preparing models for specific slicers.
-
Post-processing in the chat: the agent can retexture, remesh, split, repair, rig, and animate a model, with animations such as walking, running, or custom motions.
Create a custom 3D printing gift
Many makers know how to print but not how to model a custom object from scratch. To create a custom printed gift, follow these steps:
-
Start with a pet photo, a child's drawing, a name, a story, or a gift idea.
-
Explore custom 3D concepts with the agent.
-
Select a concept, and generate a 3D model.
-
Download the model as STL or 3MF, and then slice and print it.
Before you print, review the geometry, scale, wall thickness, and slicer settings.
Create a style-consistent set of game assets
Indie developers often need not just one asset but a set that looks like it belongs to the same world. To create a consistent set, follow these steps:
-
Describe a visual direction, for example "Low-poly medieval fantasy, warm color palette, hand-painted look."
-
Ask the agent for batches of props, characters, or environment concepts.
-
Select your favorites, ask for variations, and check that the style stays consistent.
-
Convert the concepts to 3D, and download them for your engine.
Compare 3D Agent with Text to 3D and Image to 3D
The following table compares 3D Agent with the single-step generation tools.
| Factor | 3D Agent | Text to 3D | Image to 3D |
|---|---|---|---|
| Input | Text, images, sketches, or stories | Text only | Images only |
| Multi-step process | Yes, in one continuous conversation | No, a single prompt | No, a single upload |
| Batch concept exploration | Yes, several concepts at a time | Up to 10 models per prompt with Meshy 6 | One model per image, or up to 10 images with Batch Images to 3D |
| Style consistency | Kept across the conversation | Depends on consistent prompts | Depends on the input images |
| Speed for a single model | Slower, because it includes brainstorming | Faster | Faster |
| Precision | Good for concepts | Good for well-defined prompts | Best for matching an image |
| Built-in Q&A | Yes, ask 3D questions in the chat | No | No |
Troubleshooting
The following table lists common problems, their causes, and how to fix them.
| Problem | Cause | Fix |
|---|---|---|
| The agent doesn't understand your request | The description is too vague | Say what you want to create, its style, and how you plan to use it. |
| Batch concepts aren't consistent | The style direction isn't clear enough | Give a clear art direction at the start, and name specific styles or moods. |
| The 3D model differs from the selected concept | Converting a 2D concept to 3D adds some variation | Use Image to 3D with the concept image for a closer match. |
FAQ
- What is Meshy 3D Agent?
- Meshy 3D Agent is an AI agent for 3D creation. In one conversation, it helps you brainstorm ideas, generate visual concepts in batches, ask 3D creation questions, and turn the concepts you select into 3D models.
- How is Meshy 3D Agent different from a text-to-3D tool?
- A text-to-3D tool usually generates one result from one prompt. 3D Agent supports a multi-step chat, where you explore directions, refine ideas, generate several concepts, and turn the ones you select into 3D models in one conversation.
- Is Meshy 3D Agent the same as ChatGPT for 3D?
- 3D Agent brings a conversational AI experience to 3D creation, but it isn't a general chat interface. It's built for 3D work: brainstorming, visual concept generation, 3D creation Q&A, and turning ideas into 3D models.
- Who is Meshy 3D Agent built for?
- 3D printing makers, indie game developers, 3D artists, designers, and other creators who want to move from an idea to a 3D model faster.
- Can Meshy 3D Agent be used for 3D printing?
- Yes. 3D Agent helps you create custom 3D concepts that can go into a 3D printing workflow. Before you print, review the geometry, scale, wall thickness, and slicer settings.
- Can Meshy 3D Agent generate 3D assets for games?
- Yes. 3D Agent helps indie developers and creators explore batches of 3D asset concepts in a consistent creative direction. Depending on the game, you might need to optimize, polish, or adapt the models before you use them in production.
- Can Meshy 3D Agent edit existing 3D models?
- Yes. Upload a model to the chat, or open a model from your Workspace in 3D Agent. You can then ask the agent to retexture, remesh, split, repair, rig, or animate it.
Next steps
-
To write better requests for the agent, see Prompting best practices.
-
For a game asset workflow that uses the agent, see Game assets.
-
For a print-focused workflow that uses the agent, see 3D printing workflow.