QGIS Model Context ProtocolBLOCK
Model Context Protocol (MCP) that allows LLMs to use QGIS Desktop
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
QGISMCP connects QGIS to Claude AI through the Model Context Protocol (MCP), allowing Claude to directly interact with and control QGIS. This integration enables prompt assisted project creation, layer loading, code execution and more.
This project is strongly based on the BlenderMCP project by Siddharth Ahuja
Features
- Two-way communication: Connect Claude AI to QGIS through a socket-based server.
- Project manipulation: Create, load and save projects in QGIS.
- Layer manipulation: Add and remove vector or raster layers to a project.
- Execute processing: Execute processing algorithms (Processing Toolbox).
- Code execution: Run arbitrary Python code in QGIS from Claude. Very powerful, but also be very cautious using this tool.
Components
The system consists of two main components:
- QGIS plugin: A QGIS plugin that creates a socket server within QGIS to receive and execute commands.
- MCP Server: A Python server that implements the Model Context Protocol and connects to the QGIS plugin.
Installation
Prerequisites
- QGIS 3.X (only tested on 3.22)
- Claude desktop
- Python 3.10 or newer
- uv package manager:
If you're on Mac, please install uv as
brew install uv
On Windows Powershell
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
Otherwise installation instructions are on their website: Install uv
⚠️ Do not proceed before installing UV
Download code
Download this repo to your computer. You can clone it with:
gi
cecb6c147968OBSERVED · 2026-09-26Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.
claude mcp add qgis-mcp -- uvx qgis-mcp
{
"mcpServers": {
"qgis-mcp": {
"command": "uvx",
"args": [
"qgis-mcp"
]
}
}
}Exposed tools (15)
8 read · 6 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
add_raster_layer | write | Add a raster layer to the project. |
add_vector_layer | write | Add a vector layer to the project. |
create_new_project | write | Create a new project a save it |
execute_code | write | Execute arbitrary PyQGIS code provided as a string. |
execute_processing | write | Execute a processing algorithm with the given parameters. |
get_layer_features | read | Retrieve features from a vector layer with an optional limit. |
get_layers | read | Retrieve all layers in the current project. |
get_project_info | read | Get current project information |
get_qgis_info | read | Get QGIS information |
load_project | read | Load a QGIS project from the specified path. |
ping | read | Simple ping command to check server connectivity |
remove_layer | destructive | Remove a layer from the project by its ID. |
render_map | read | Render the current map view to an image file with the specified dimensions. |
save_project | write | Save the current project to the given path, or to the current project path if not specified. |
zoom_to_layer | read | Zoom to the extent of a specified layer. |
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- declared (2 observation(s))
- Shell
- declared (1 observation(s))
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (3)
exec(code, namespace)
remove_layer
Gates applied: no_behavioural_pass, no_license.
cecb6c147968full audit observations/trust-audit/mcp-server/jjsantos01__qgis-model-context-protocol.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-09-26 | cecb6c147968 | BLOCK | D | 69 | first audit |
Questions
What is the QGIS Model Context Protocol MCP server?
Model Context Protocol (MCP) that allows LLMs to use QGIS Desktop
What tools does QGIS Model Context Protocol expose?
15 in total: 8 read-only, 6 that write, and 1 that can delete or overwrite (remove_layer). Every one is listed on this page with its risk.
Is QGIS Model Context Protocol safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does QGIS Model Context Protocol need?
No credential environment variables were found in its source, so it appears to need none.
How current is this page?
The grade is for one exact copy of the source (cecb6c147968), read on 2026-09-26. The repository is watched and re-audited when it changes.