# Meshy (3D AI) MCP > Meshy (3D AI) connects generative artificial intelligence directly to your agent, turning simple text prompts or reference images into professional-grade, textured 3D meshes. Generate full assets, refine topology, and create animations—all without leaving your development environment. It bridges the gap between a concept sketch and a game-ready model. ## Overview - **Category:** design-creative - **Price:** Free - **Tags:** 3d-modeling, generative-ai, text-to-3d, pbr-textures, mesh-generation, asset-creation ## Description Meshy brings high-quality 3D asset creation into your AI workflow. You no longer have to switch between dozens of specialized software programs just to get started on a new piece of concept art or geometry. Instead, you describe what you need—whether that's 'a rusted cyberpunk motorcycle' or 'three reference photos of a medieval shield'—and Meshy handles the heavy lifting. It generates initial 3D previews from text and builds detailed models from multiple images, automatically applying PBR maps so they look realistic right out of the gate. If you need to adjust the geometry for a game engine, you can run remeshing tools to fix the topology or apply new styles using retexturing techniques while keeping the original shape. Need to manage your work? You track all generation tasks in one spot. Integrating this MCP via Vinkius means you connect once from any AI client and immediately gain access to professional-grade 3D visualization capabilities, speeding up everything from prototyping game assets to visualizing architectural concepts. ## Tools ### create_text_to_image Generates a 2D image based on a text prompt. ### delete_text_to_3d_task Removes an existing Text to 3D generation task by its ID. ### get_balance Retrieves the current account balance for usage tracking. ### get_text_to_3d_task Fetches the status and details of a specific Text to 3D task using its unique ID. ### list_text_to_3d_tasks Retrieves a list of all past or current Text to 3D generation tasks. ### repair_printability Analyzes and repairs a model mesh, making it suitable for physical 3D printing processes. ### analyze_printability Checks the current status of a 3D model to determine if it can be successfully printed. ### create_animation Generates an animation sequence for a given 3D asset. ### create_image_to_3d Turns one or more reference images into a detailed, structured 3D object. ### create_image_to_image Transforms an input image while maintaining its structure and composition. ### create_multi_color_print Prepares a model for printing that requires multiple distinct colors. ### create_multi_image_to_3d Builds a 3D model using several reference images, such as views from different angles. ### create_remesh Adjusts the internal geometry (topology) of an existing mesh to fit new requirements. ### create_retexture Applies entirely new materials or styles to a 3D model using text or image guidance. ### create_rigging Creates the necessary skeletal structure for a character or object so it can be posed and animated. ### create_text_to_3d_preview Initiates the first step of generating an initial 3D model from a text description. ### create_text_to_3d_refine Refines and improves the geometry and details of a preliminary 3D preview mesh. ## Prompt Examples **Prompt:** ``` Create a 3D preview of a futuristic cyberpunk motorcycle. ``` **Response:** ``` I've initiated the `create_text_to_3d_preview` task for your cyberpunk motorcycle. The task ID is 'msy_abc123'. I'll let you know as soon as the preview mesh is ready. ``` **Prompt:** ``` Generate a 3D model from this image: https://example.com/character.png ``` **Response:** ``` Starting the `create_image_to_3d` process. I'm uploading the reference image to Meshy's AI to generate the geometry and textures. This usually takes a few minutes. ``` **Prompt:** ``` List my recent 3D generation tasks. ``` **Response:** ``` Fetching your history via `list_text_to_3d_tasks`... I found 3 recent models: a 'Sci-fi Crate' (Succeeded), a 'Stone Arch' (Succeeded), and a 'Dragon' (In Progress). ``` ## Capabilities ### Generate 3D models from text Input a descriptive prompt to generate an initial 3D preview mesh. ### Model objects from images Upload single or multiple reference photos and convert them into structured, detailed 3D geometry. ### Adjust model topology Run remeshing tools to change the triangle structure of a mesh, targeting specific polycounts for various applications. ### Change materials and styles Apply new textures or entire visual styles to an existing 3D model using text or image guidance without losing its core shape. ### Create animations Generate sequential frames for a 3D object, turning static geometry into movement. ## Use Cases ### Conceptualizing a New Game Prop A game developer needs a 'glowing crystal shard.' Instead of sketching it and modeling it, they ask their agent to use `create_text_to_3d_preview`. The resulting mesh is fast enough for immediate gameplay testing. They then run `create_retexture` to give it the specific glowing effect needed. ### Turning Product Photos into Assets A design team has 4 photos of a futuristic piece of furniture from different angles. They use their agent with `create_multi_image_to_3d` to generate a full, usable 3D model that they can then export for client visualization. ### Preparing Art for Physical Prints An artist designs an intricate mask. Before sending it to the printer, they run `analyze_printability` and use `repair_printability`. The agent flags structural weak points and fixes them, ensuring a successful physical print. ### Improving Character Models A character model was generated but has messy geometry. The 3D artist uses the agent to run `create_remesh` on the problematic area, adjusting the topology so it's clean enough for rigging and animation. ## Benefits - Speed up prototyping. Instead of spending hours on initial block-outs, you can use `create_text_to_3d_preview` to generate a basic mesh in minutes, letting your agent test the concept immediately. - Overcome source material limitations. If you only have photos, don't worry about perfect angles; simply run `create_multi_image_to_3d` to build a full 3D object from multiple references. - Fix technical assets on the fly. Need a model with cleaner geometry for a game engine? Use the `create_remesh` tool to adjust the topology and hit your target polycount without starting over. - Maintain integrity while styling. You can radically change a model's look—say, from stone to chrome—using `create_retexture`, but the underlying structure remains intact. - Streamline complex animation. Generating movement used to require multiple artists; now you can use `create_animation` to bring static geometry to life through your agent. ## How It Works The bottom line is, Meshy automates the complex initial stages of 3D asset pipelines so your agent can handle it conversationally. 1. Subscribe to this MCP and provide your Meshy API Key. 2. Call the relevant tool using your AI agent (e.g., 'create_image_to_3d') within your preferred client. 3. Receive a task ID for the generation, which you then use to check on progress or refine the final asset. ## Frequently Asked Questions **How do I start generating 3D models with Meshy (3D AI)?** You first initiate a preview using `create_text_to_3d_preview`. This generates the basic mesh, and you can then use `create_text_to_3d_refine` to improve it further. **Can Meshy (3D AI) create models from pictures?** Yes. You have two main options: use `create_image_to_3d` for a single reference photo, or use `create_multi_image_to_3d` if you have several photos taken from different angles. **What is the difference between Meshy (3D AI) and standard 3D software?** Standard software requires manual input for every step. This MCP automates that initial pipeline, allowing your agent to handle complex tasks like remeshing or rigging with a single tool call. **Does Meshy (3D AI) help with animation?** Yes, you can use the `create_animation` tool. This generates sequential frames for your 3D asset, turning static geometry into movement. **How do I make my model ready for printing using Meshy (3D AI)?** You first run `analyze_printability` to check the mesh. If issues are found, use `repair_printability` to fix it before proceeding with multi-color printing via `create_multi_color_print`.