Describe your goal and OpenGeodex orchestrates local data collection, processing, analysis, and cartography — turning complex GIS work into traceable, reusable agent tasks.
Coordinator for Q&A and dispatch, collector for acquisition, processor for analysis, with multi-turn planning.
Supports OpenAI-compatible providers such as GLM and DeepSeek, with multiple models configured and switched on demand.
DEM, imagery, land cover, admin boundaries, place names, population, weather, OSM, coordinate correction, and geocoding.
Vector and raster processing, contours, PostGIS / GeoServer access, SQL safety, map styling, and output.
Built-in, plugin, and user-defined skills; declarative plugins connect GDAL, GMT, QGIS, and more geoscience tools.
Everything runs locally. The LLM sees key metadata only — never your real geographic data.
Approval for risky operations, sandboxed Python / Node / Shell / PowerShell execution, and filesystem safety.
OCR, speech synthesis, stealth web crawl, web search, plus Feishu / WeCom async notifications and remote commands.
Windows release packages are available now. macOS and Linux are coming soon.
Deliverable: PNG / PDF topographic map
Got it. I will collect DEM, calculate slope, aspect, and hillshade, generate contours, and export the layout.
Deliverable: choropleth map
I will collect statistics, join them to admin boundaries, classify values, and produce the thematic map.
Deliverable: candidate-site comparison map
I will collect POI and roads, run buffer and density analysis, then style a candidate-site comparison map.
Deliverable: ingest complete + QA report
I will run GDAL/OGR reprojection, ingest to PostGIS, create indexes, and return a QA report.
