MCP Chatting

Example Use Cases

• 1

  Answer user questions about QGIS functions and workflows by searching and retrieving relevant QGIS documentation.

• 2

  A local, self-hosted chat interface to query LLMs with custom API keys and avoid third-party chat platform limits.

• 3

  Combine LLM reasoning with external QGIS knowledge (RAG) when building GIS-focused assistants or plugins.

Description

mcpchatting is a hackathon project that integrates a Model Context Protocol (MCP) server with a built-in chat client, allowing users to run LLMs against QGIS documentation via Retrieval-Augmented Generation (RAG). It supports multiple language model backends (e.g., OpenAI, Deepseek) and lets users supply their own API keys to avoid reliance on external hosted chat platforms. The system includes a QGIS MCP plugin and a local client to start and query the MCP server, enabling more accurate, context-aware responses about QGIS functions and docs.