qgis-mcp-codex
Connects OpenAI Codex Desktop to QGIS Desktop for natural language GIS operations, including project management, layer manipulation, and algorithm execution.
README
OpenAI Codex QGIS MCP
OpenAI Codex QGIS MCP connects Codex Desktop to QGIS Desktop through the Model Context Protocol (MCP). It bundles both sides needed for Codex:
- A Codex-compatible MCP server.
- A QGIS Desktop plugin that opens a local socket server.
- A Codex plugin manifest, MCP config, and GIS pre-flight skill.
This project is intentionally separated from the Claude-compatible QGIS MCP implementation. The Codex bundle uses port 9877 by default so it can coexist with the original Claude-oriented QGIS MCP plugin on port 9876.
Features
- Connect OpenAI Codex Desktop to QGIS Desktop.
- Create, load, inspect, and save QGIS projects.
- Add and remove vector or raster layers.
- List layers and inspect vector features.
- Zoom to layers and render the current map view to an image.
- Run QGIS Processing algorithms.
- Execute focused PyQGIS code inside QGIS Desktop when needed.
- Provide a Codex GIS skill for CRS, geometry, schema, extent, and null-geometry pre-flight checks.
Repository Layout
src/qgis_mcp_codex/- bundled Codex-compatible MCP server and socket client.qgis_mcp_codex_plugin/- QGIS Desktop plugin for Codex, listening onlocalhost:9877.plugins/qgis-mcp-codex/- Codex plugin manifest, MCP config, launcher, and skill..agents/plugins/marketplace.json- local Codex plugin marketplace entry.scripts/package_qgis_plugin.ps1- helper to package the QGIS plugin as a ZIP.
Requirements
- QGIS 3.x.
- OpenAI Codex Desktop.
- Python 3.12 or newer for the MCP server.
uvpackage manager.- Windows PowerShell for the included packaging helper.
Install uv on Windows PowerShell:
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
QGIS Desktop Plugin Setup
Manual install:
-
Copy the folder
qgis_mcp_codex_pluginto your QGIS profile plugin directory. -
The final path usually looks like:
C:\Users\User\AppData\Roaming\QGIS\QGIS3\profiles\default\python\plugins\qgis_mcp_codex_plugin -
Confirm these files are directly inside that folder:
metadata.txt __init__.py qgis_mcp_codex_plugin.py -
Restart QGIS Desktop.
-
Open
Plugins > Manage and Install Plugins. -
Enable
QGIS MCP Codex. -
Open
Plugins > QGIS MCP Codex > QGIS MCP Codex. -
Click
Start Server.
ZIP install:
- Open PowerShell in the repository root, the folder that contains this
README.md. - Run:
.\scripts\package_qgis_plugin.ps1
- This creates:
dist\qgis_mcp_codex_plugin.zip
-
In QGIS, open
Plugins > Manage and Install Plugins > Install from ZIP. -
Select the generated ZIP file:
<repo-root>\dist\qgis_mcp_codex_plugin.zip -
Click
Install Plugin, restart QGIS if needed, then enableQGIS MCP Codex.
Codex Desktop Setup
Edit your Codex Desktop config file. On Windows, the usual path is:
C:\Users\<your-user>\.codex\config.toml
For example, on a machine where the Windows user is User, the path is:
C:\Users\User\.codex\config.toml
If your Codex config lives somewhere else, edit that config.toml instead.
Add this local plugin marketplace configuration:
[marketplaces.qgis-mcp-local]
source_type = "local"
source = 'C:\path\to\OpenAICodex-QGIS-MCP'
[plugins."qgis-mcp-codex@qgis-mcp-local"]
enabled = true
Replace C:\path\to\OpenAICodex-QGIS-MCP with the absolute path to this repository.
Restart Codex Desktop after editing the config.
The Codex MCP server entry is named qgis. It launches:
plugins/qgis-mcp-codex/scripts/launch_qgis_mcp.py
The launcher runs the bundled server:
src/qgis_mcp_codex/qgis_mcp_server.py
Test The Connection
-
Start QGIS Desktop.
-
Start the
QGIS MCP Codexplugin server on port9877. -
Restart Codex Desktop if you just changed plugin config.
-
Ask Codex:
Use the QGIS MCP tool to ping QGIS.
Expected response includes pong: true.
You can also ask:
Use the QGIS MCP tool to get QGIS info and current project info.
Available MCP Tools
pingget_qgis_infoload_projectcreate_new_projectget_project_infoadd_vector_layeradd_raster_layerget_layersremove_layerzoom_to_layerget_layer_featuresexecute_processingsave_projectrender_mapexecute_code
execute_code runs arbitrary PyQGIS code inside QGIS Desktop. Use it carefully and prefer narrower tools or QGIS Processing algorithms when possible.
Troubleshooting
- If Codex cannot connect, confirm QGIS Desktop is open and
QGIS MCP Codexis running on port9877. - If QGIS does not detect the plugin, confirm the folder is under
python\pluginsand is not nested twice. - If port
9877is busy, change the port in the QGIS plugin UI and update the bundled MCP server default insrc/qgis_mcp_codex/qgis_mcp_server.py. - If
uvis not found, install it and restart Codex Desktop. - If Codex does not show the plugin after config changes, restart Codex Desktop.
Credits
This project is based on and adapted from jjsantos01/qgis_mcp, which connects QGIS Desktop to Claude Desktop through MCP.
The original QGIS MCP project is itself based on the BlenderMCP approach by Siddharth Ahuja. This repository separates the Codex-compatible implementation into its own MCP server, QGIS plugin, Codex plugin metadata, and default port.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.