GeoAgent: An Agent-Based QGIS Plugin for Natural-Language-Driven Geospatial Analysis with Multi-Backend Large Language Model Support
Abstract
Geospatial analysis is essential across the earth and environmental sciences, yet conventional geographic information systems (GIS) rely on a procedure-oriented paradigm that demands substantial operational expertise. We present GeoAgent, an open-source QGIS plugin that enables goal-driven geospatial workflows through natural language conversation powered by large language models (LLMs). GeoAgent features two operating modes: a General mode for exploratory querying and layer management, and a Processing mode for multi-step workflow execution. Built on LangGraph state machines, the system decomposes requests into dependency-ordered task queues, discovers algorithms across the entire QGIS Processing registry (~250+ native, GDAL, GRASS GIS, and SAGA tools) via lexical scoring on LLM-generated keywords, introspects parameter schemas at runtime, and executes workflows with LLM-steered error diagnosis and retry loops. Schema-validated structured outputs ensure consistent performance across interchangeable cloud (OpenAI, Google Gemini, Anthropic) and local privacy-preserving (Ollama) backends. We demonstrate GeoAgent through a reproducible flood-exposure screening case study using a bundled dataset and compare its architecture against existing LLM-GIS integrations. GeoAgent is released under the MIT license on the official QGIS Plugin Repository.
Related
- An Artificial Intelligence Integrated Geospatial Decision-Support Framework for Scenario-Based Climate-Adaptive Land-Use Optimization and Environmental Risk Reduction
- GeoAgent: Contract-Constrained LLM Agents for Seismic Data Processing
- Post-processing Subglacial Physics Model with Per-site JEPA Specialist Models Enable Natural-language Glaciological Intelligence
- High-performance geodynamics on GPUs using the PETSc CUDA backend (GAUZZ v1.0.0)
- A geospatial and community perception analysis of the impacts of land use and land cover change on ecosystem service values
BibTeX
@misc{kshetri2026geoagent,
title = {{GeoAgent: An Agent-Based QGIS Plugin for Natural-Language-Driven Geospatial Analysis with Multi-Backend Large Language Model Support}},
author = {Tek Bahadur Kshetri and Rabin Ojha},
howpublished = {EarthArXiv},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23048185},
url = {https://eartharxiv.org/repository/view/15304/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - GeoAgent: An Agent-Based QGIS Plugin for Natural-Language-Driven Geospatial Analysis with Multi-Backend Large Language Model Support
AU - Tek Bahadur Kshetri
AU - Rabin Ojha
PY - 2026
DA - 2026/10/01
PB - EarthArXiv
DO - 10.5281/zenodo.23048185
UR - https://eartharxiv.org/repository/view/15304/
AB - Geospatial analysis is essential across the earth and environmental sciences, yet conventional geographic information systems (GIS) rely on a procedure-oriented paradigm that demands substantial operational expertise. We present GeoAgent, an open-source QGIS plugin that enables goal-driven geospatial workflows through natural language conversation powered by large language models (LLMs). GeoAgent features two operating modes: a General mode for exploratory querying and layer management, and a Processing mode for multi-step workflow execution. Built on LangGraph state machines, the system decomposes requests into dependency-ordered task queues, discovers algorithms across the entire QGIS Processing registry (~250+ native, GDAL, GRASS GIS, and SAGA tools) via lexical scoring on LLM-generated keywords, introspects parameter schemas at runtime, and executes workflows with LLM-steered error diagnosis and retry loops. Schema-validated structured outputs ensure consistent performance across interchangeable cloud (OpenAI, Google Gemini, Anthropic) and local privacy-preserving (Ollama) backends. We demonstrate GeoAgent through a reproducible flood-exposure screening case study using a bundled dataset and compare its architecture against existing LLM-GIS integrations. GeoAgent is released under the MIT license on the official QGIS Plugin Repository.
ER -