AI integration

Connect AI agents and coding assistants to Meshy, with the Meshy MCP server for tool calls, or with llms.txt and llms-full.txt for reading the docs.

If you use Claude Code, Cursor, Windsurf, Codex, or another MCP client, install the MCP server. If your chat agent doesn't support MCP, point it at https://docs.meshy.ai/llms.txt instead.

What is MCP

The Model Context Protocol (MCP) is an open standard that lets AI assistants call external tools and get structured context. The Meshy MCP server wraps the Meshy REST API as tools that your agent can call. With them, the agent can generate a model, check a task, and download the result without hand-written HTTP code.

The MCP server is open source. It's published on npm as @meshy-ai/meshy-mcp-server, and its source is in the meshy-mcp-server repository on GitHub.

You need a Meshy API key. Create one on the API Keys page of the Developer Platform. The MCP server reads it from the MESHY_API_KEY environment variable.

Install the MCP server

To install the server, use the tab for your AI client. Replace msy_YOUR_API_KEY with your API key.

# Detects installed AI clients and configures Meshy for each
npx add-mcp @meshy-ai/meshy-mcp-server --env MESHY_API_KEY=msy_YOUR_API_KEY

npx skills add meshy-dev/meshy-3d-agent

Available tools

The MCP server offers the Meshy REST API as the following tools, grouped by capability.

3D generation

• meshy_text_to_3d: create an untextured preview mesh from a text prompt
• meshy_image_to_3d: create a 3D model from one image
• meshy_multi_image_to_3d: create a 3D model from multiple images of the same object
• meshy_text_to_3d_refine: add texture to a preview mesh

Post-processing

• meshy_remesh: change the topology or polycount of an existing model, or both
• meshy_retexture: apply a new texture to an existing model
• meshy_rig: add a skeleton to a 3D humanoid character
• meshy_animate: apply an animation to a rigged character
• meshy_convert: convert a model to another file format
• meshy_resize: resize a model to real-world dimensions
• meshy_uv_unwrap: generate a new UV layout for a model
• meshy_text_to_motion: generate an animation from a text prompt
• meshy_list_animations: look up the preset animation action IDs, for free

Creative Lab

• meshy_creative_lab: turn a photo into a finished Creative Lab product (runs the prototype and build tasks)

Image generation

• meshy_text_to_image: create a 2D image from text
• meshy_image_to_image: create a 2D image from a reference image

Task management

• meshy_get_task_status: check task status and download URLs
• meshy_list_tasks: list recent tasks, optionally filtered by type or status
• meshy_cancel_task: delete a task. Meshy removes and refunds a PENDING task, and deletes a finished task. An IN_PROGRESS task can't be canceled, and the API returns 409.
• meshy_download_model: download a finished model file and save it locally

Workspace

• meshy_list_models: list your recent Text to 3D models. For other task types, use meshy_list_tasks.

3D printing

• meshy_send_to_slicer: find the slicers installed on your computer, and return a command that opens the model in one. It runs locally and doesn't call the Meshy API.
• meshy_analyze_printability: check a model's FDM printability for free, including watertightness, volume, holes, non-manifold edges, and degenerate faces
• meshy_process_multicolor: convert a textured model into a multi-color 3MF file for printing
• meshy_repair_printability: repair a model for FDM printing
• meshy_auto_split: split a model into separately printable parts

Account

• meshy_check_balance: check your remaining credits
• meshy_list_usage: list your team's finished API tasks and the credits each one used, for Studio and Enterprise teams

Example prompts

To start, try these prompts in your MCP-enabled chat:

Generate a 3D fox from the prompt "a cartoon fox sitting", preview it, then
texture it with PBR maps. Download the final GLB to ./outputs.

Take the image at https://example.com/sculpture.jpg and convert it into a
riggable 3D character. Use `should_remesh: true` and 50k target polycount.

What's my current Meshy credit balance?

List my last 10 successful text-to-3d tasks and download the top 3 as GLB
into ./downloads/.

llms.txt and llms-full.txt

If your agent doesn't support MCP, or you want to put the Meshy docs into a prompt directly, point it at the plain-text versions of the docs:

• llms.txt: a compact index with integration instructions, such as how to poll async tasks, authentication, rate limits, model choice, and common mistakes.
• llms-full.txt: every API page in one file, so an agent can read the docs in one request.
• Per-page Markdown: add .md to any endpoint URL, for example https://docs.meshy.ai/api/text-to-3d.md.

Meshy regenerates all three every time the docs site builds, so they match the HTML docs.

FAQ

Is the MCP server stateless?

It doesn't store your credentials. The server keeps your MESHY_API_KEY in memory and never writes it to disk. By default, it saves the models it downloads, with their metadata, in a meshy_output folder in its working directory. To save a download somewhere else, give the download tool a save_to path.

Does MCP cost the same as the REST API?

Yes. MCP tool calls make the same REST calls and use credits at the same rate. See the Pricing page for every price.

What are the rate limits?

The MCP server's calls count toward the same rate limits as your REST API calls. See the Rate limits page for the limits of each plan.

What Meshy data can the MCP server access?

Only what your MESHY_API_KEY can access through the REST API, with the same permissions.

How do I report issues?

Open an issue in the meshy-mcp-server repository on GitHub: https://github.com/meshy-dev/meshy-mcp-server/issues