AETumi MCP is a hosted Model Context Protocol server. Connect it once and your AI coding assistant can discover production-ready Three.js, WebGL and React Three Fiber components, 3D website templates and interactive scenes from the AETumi library — without leaving the editor.
Requirements
- An MCP-compatible AI coding client: Claude Code, Cursor or Codex.
- Internet access to the hosted endpoint. Nothing to install or build locally.
Hosted endpoint
The server runs over Streamable HTTP (JSON-RPC 2.0) at:
https://mcp.aetumi.app
Connect your client
Claude Code
claude mcp add --transport http aetumi https://mcp.aetumi.app
Cursor / Codex & other MCP clients — add a Streamable HTTP MCP server pointing at https://mcp.aetumi.app in your client's MCP configuration.
Test the connection
List the available tools; you should see ten read-only tools (six catalog, four experience-discovery). In Claude Code, the server appears as aetumi. A quick raw check:
curl -s -X POST https://mcp.aetumi.app -H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Your first search
Just ask in natural language — the assistant chooses the right tool:
- “Find me a Three.js hero section for a luxury product website.”
- “Show me 3D web experiences for automotive.” (uses
filter_by_industry) - “Recommend a build for a 3D headphone website.” (uses
recommend_3d_web_stack)
Each result includes a live interactive reference and the id you can pass to get_3d_web_asset for full detail.
Next steps
See the full MCP Tools Reference for every tool, its arguments and example queries. Explore the AETumi MCP overview, or read the source on GitHub.

