Sharing cesium-mcp: a community MCP integration for CesiumJS

General
1

Hi everyone!

I’ve been really excited to see CesiumGS’s official work on cesium-ai-integrations — it’s great to see MCP support coming to the Cesium ecosystem.

Inspired by that direction, I’d like to share a community project I’ve been working on: cesium-mcp. It’s an open-source MCP server that lets AI assistants (Claude, Copilot, Cursor, etc.) interact with a CesiumJS Viewer through natural language.

What it does

cesium-mcp provides a browser-side SDK (cesium-mcp-bridge) that connects to a Node.js MCP server (cesium-mcp-runtime) via WebSocket. This allows AI agents to execute CesiumJS operations like:

  • Camera navigation (flyTo, setView, orbit)

  • Adding entities (markers, polygons, models, labels, and more)

  • Managing layers (GeoJSON, 3D Tiles, imagery services)

  • Creating trajectory animations

  • Taking screenshots

Currently it has 43 tools organized in 10 groups, with a dynamic discovery mechanism so the AI can activate tool groups on demand.

How it relates to the official project

cesium-mcp started independently before the official cesium-ai-integrations was announced. Since then, I’ve been aligning the tool design with the official approach — the camera, entity, and animation features follow similar patterns to the official servers.

I see this as complementary to the official effort — exploring additional use cases like layer management, GeoJSON loading, and remote deployment (via Streamable HTTP) that might be useful for the community. I also submitted a feature request (Issue #19) with some ideas for the official project.

Quick start


# In your CesiumJS project

npm install cesium-mcp-bridge

# In your app code

const bridge = new CesiumMcpBridge(viewer, { port: 9100 });

bridge.connect();

# Start the MCP server

npx cesium-mcp-runtime

Links

I’d love to hear any feedback or suggestions from the community. Especially interested in what use cases you’d find most valuable for AI + CesiumJS integration.

Thanks!

2 Likes