Meshy MCP
Generate 3D inside your agent
One URL or one command connects Meshy to Claude, ChatGPT, Cursor, Replit, OpenClaw, and any MCP client.
Set up in one command
Pick your client, then either paste one line into your agent and let it do the setup, or copy the config yourself.
1
Add the server
Already in a session with Claude Code, Cursor or Codex? Paste this and it will handle the setup.
text
Install the Meshy MCP server for me. Docs: https://github.com/meshy-dev/meshy-mcp-server Use this env var: MESHY_API_KEY=msy_YOUR_API_KEY
2
Activate and verify
Click Add, then Connect, and approve in the pop-up — then ask Claude for the model you want.
Verify with
@ meshyThen just ask for the model you want.
Paste this into your prompt and the agent reads the full tool reference itself.
One conversation, a finished asset
The same chat works in every client you connected above — you describe, the agent runs the Meshy tools, you approve.
What your agent can do now
One connection puts 24 Meshy tools in the conversation — generation is only the start.
Batch out whole sets
One prompt, a full prop set — crates, signs, baskets and chests, ready for your level.
Turn anything into 3D
Text, photos and sketches become textured meshes — or clean sculpts for printing.
It knows 3D
Formats, printability, broken meshes — the tools carry the know-how, so the agent answers and fixes instead of guessing.
24Meshy tools
One connection, and every one of them is callable by name.
Built for Every 3D Workflow
From a single chat, the agent slots into the pipeline you already run — games, 3D printing, and product design — exporting the right format for each.
>Make a chibi ice dragon game character, cartoon style, and bake 2K textures.
●meshy_text_to_3d
Frequently asked questions
Model Context Protocol is an open standard for connecting AI agents to external tools and data. An MCP server exposes a set of tools, and any MCP-compatible client can call them. Meshy publishes one for 3D generation.
Claude, ChatGPT, Cursor, Replit and OpenClaw have step-by-step setup on this page, with Claude Code, Codex, Windsurf, VS Code, Claude Desktop, Zed, Gemini CLI and Warp in the More menu — plus a universal config for anything else that speaks MCP.
Yes. The server authenticates with a Meshy API key, which requires a Pro plan or above. Create one in your account settings and pass it as MESHY_API_KEY.
Tool calls consume the same credits as the equivalent REST API calls — nothing costs more because an agent made the request. Printability analysis is free, convert and resize cost 1 credit each, UV unwrap costs 5, and topology repair and multi-color 3MF cost 10. Generation tools cost the most, so use meshy_check_balance to see what is left before a long chain.
The plugins put Meshy inside a specific 3D application, driven by you clicking. The MCP server puts Meshy inside an AI agent, driven by the agent deciding. Plenty of people run both.
About a minute for a textured mesh from a text prompt. Rigging, animation and conversion are quicker. Tools return a task id immediately, so the agent polls with meshy_get_task_status instead of blocking.
Cursor and VS Code need a manual toggle after install — see the activate step for your client above. If the tools still do not appear, check that mcp.json is valid JSON with no trailing commas, and that MESHY_API_KEY sits inside an env block rather than in args.
Yes. If npx fails to spawn, wrap the command with cmd /c. On Claude Code for Windows, use claude mcp add-json rather than the flag form.
Commercial rights follow your Meshy plan, not the interface you generated through. Paid plans include commercial licensing.
The server runs locally with your own API key; a hosted endpoint is on the roadmap. For shell scripts and CI, install meshy-cli from npm — it signs in through your browser instead.
Name the server meshy — every config on this page does — and say so in the prompt: “use Meshy to…”. Agents route by tool descriptions, and naming the product is the most reliable disambiguation.
Every asset also lands in your Meshy workspace — a model generated from an agent can be retextured, edited and exported on the web later, and the other way around.
Each member uses their own API key, so usage and permissions stay per-person. A one-config-for-the-whole-org setup depends on the hosted server on our roadmap.
Explore More Features
3D Agent
Chat from an idea, photo, or sketch to a print-ready 3D model
Image to 3D
Turn any photo into a detailed 3D model in seconds
Text to 3D
Describe it, generate it - 3D models from text prompts
AI Texturing
Add realistic PBR textures to any 3D model with AI
AI Image Generator
Generate multi-view images optimized for 3D creation
Animation
Auto-rig and animate any 3D character in seconds
Put 3D generation in your agent
Create an API key and connect in one command.