# SSWorld MCP server

> Lets AI agents author, compile, render, and inspect 3D geographic scenes using WebGPU with ray-traced global illumination.

- Repository: https://github.com/DataArche/SSWorld
- Page: https://hostedmcp.org/servers/dataarche/ssworld
- Transport: —
- Stars: 5
- License: Apache-2.0

## Tools (13)

- `ssworld_catalog`
- `ssworld_project_list`
- `ssworld_project_create`
- `ssworld_source_patch`
- `ssworld_source_batch`
- `ssworld_scene_inspect`
- `ssworld_compile`
- `ssworld_preview`
- `ssworld_capture_frame`
- `ssworld_environment_read`
- `ssworld_geo_read`
- `ssworld_geometry_read`
- `ssworld_engine_status`

## FAQ

### What can an agent do with SSWorld?

An agent can author, compile, render, and inspect 3D geographic scenes written in SSDL on the SSEngine WebGPU runtime.

### How does the agent know if the render looks correct?

The agent can use the capture frame tool to receive luma percentiles, exposure tails, color coverage, and camera pose data to verify the visual output.

### Does SSWorld support real-world geographic data?

Yes, scenes can be anchored by longitude and latitude and include terrain, imagery layers, 3D Tiles, and GeoJSON.

### What are the system requirements for installation?

You need Node.js version 22 or higher, npm, and a WebGPU-compatible browser like Chrome or Edge to view previews.