Wraps the python-geoservercloud library to expose 80+ GeoServer REST API operations through MCP. You can manage workspaces, datastores, layers, and styles using natural language instead of writing API calls. Works with Claude Desktop, VS Code, Cursor, and other MCP clients. The underlying library supports the full GeoServer Cloud REST API, so you get things like PostGIS datastore connections, layer publishing, and style management. Reach for this when you're administering GeoServer instances and want to skip the web UI or manual curl commands. Configuration takes three environment variables: your GeoServer URL, username, and password.

An MCP server that wraps the python-geoservercloud library, exposing 80+ GeoServer operations as natural-language tools for AI assistants like Claude, VS Code Copilot, and other MCP-compatible clients.
Once connected, you can just ask:
test_data"roads table from my PostGIS database"One command — no manual install, uvx fetches and runs the server for you.
Simplest — no credentials up front:
claude mcp add geoserver -- uvx geoservercloud-mcp
The AI will ask you for the GeoServer URL, username, and password when it first needs them. Great for trying it out or switching between servers.
Or set the connection up front:
claude mcp add geoserver \
--env GEOSERVER_URL=http://localhost:8080/geoserver \
--env GEOSERVER_USER=admin \
--env GEOSERVER_PASSWORD=geoserver \
-- uvx geoservercloud-mcp
That's it — start Claude Code and ask it to list your workspaces to confirm it's connected.
Manage it later with claude mcp list, claude mcp get geoserver, or
claude mcp remove geoserver. Add --scope user to the add command to make it
available in every project instead of just this one.
Add this to your config file, then restart Claude Desktop:
~/Library/Application Support/Claude/claude_desktop_config.json~/.config/Claude/claude_desktop_config.json{
"mcpServers": {
"geoserver": {
"command": "uvx",
"args": ["geoservercloud-mcp"],
"env": {
"GEOSERVER_URL": "http://localhost:8080/geoserver",
"GEOSERVER_USER": "admin",
"GEOSERVER_PASSWORD": "geoserver"
}
}
}
}
Add this to .vscode/mcp.json:
{
"servers": {
"geoserver": {
"command": "uvx",
"args": ["geoservercloud-mcp"],
"env": {
"GEOSERVER_URL": "http://localhost:8080/geoserver",
"GEOSERVER_USER": "admin",
"GEOSERVER_PASSWORD": "geoserver"
}
}
}
}
# with pip
pip install geoservercloud-mcp
geoservercloud-mcp
# or run without installing (requires uv: https://docs.astral.sh/uv/)
uvx geoservercloud-mcp
Published to the MCP Registry as
io.github.ronitjadhav/geoservercloud-mcp.
| Variable | Default | Description |
|---|---|---|
GEOSERVER_URL | http://localhost:8080/geoserver | GeoServer base URL |
GEOSERVER_USER | admin | GeoServer username |
GEOSERVER_PASSWORD | geoserver | GeoServer password |
All three are optional — if you skip them, you can configure the connection at runtime by asking the AI.
Want to run it from source or contribute?
git clone https://github.com/ronitjadhav/geoservercloud-mcp.git
cd geoservercloud-mcp
poetry install # set up the environment
poetry run pytest # run the tests
poetry run geoservercloud-mcp # run the server (stdio)
Need a GeoServer to test against? cd docker && docker compose up -d spins up
GeoServer + PostGIS + the MCP server.
For the full workflow — adding new tools, linting, releasing, and how publishing works — see the Developer Guide.
This server is built on the python-geoservercloud library. For programmatic access without MCP:
from geoservercloud import GeoServerCloud
geoserver = GeoServerCloud(
url="http://localhost:8080/geoserver",
user="admin",
password="geoserver",
)
geoserver.create_workspace("my_workspace")
Full library docs: https://camptocamp.github.io/python-geoservercloud/
GEOSERVER_URLGeoServer base URL (default: http://localhost:8080/geoserver)
GEOSERVER_USERGeoServer username (default: admin)
GEOSERVER_PASSWORDsecretGeoServer password (default: geoserver)