GeoSQL: Turning Claude and Copilot into Geospatial Analytics Agents
Geosql: A Claude/Codex skill for geospatial data
I built GeoSQL to transform Claude, Codex, and GitHub Copilot into powerful geospatial agents for PostGIS, BigQuery, and Snowflake. By integrating a map-in-the-loop feedback system, this tool achieves a 4x performance boost on spatial tasks. It runs entirely locally or self-hosted, ensuring your data stays secure while enabling complex analyses like site selection and infrastructure planning without needing a SaaS account.
Without maps, the agent's text-only validation misses geometry-class errors, but adding the rendered map as a tool call lets the agent see those mistakes and self-correct.
- OtherShrezzing
>4x improvement on geospatial tasks with map in the loop.
The graph shows a baseline 2% task success rate improving to to 8% task success rate, but the evals section details 100% success rates across the board.
I'm not sure what the effectiveness of this skill is from the readme. Is it 8% success, or 100% success?
- cpa
This space is very active. I work at the French mapping agency, and we're currently building MCPs to work with our data.
See: https://github.com/ignfab/geocontext (French)
Beta MCP instance: https://geollm.beta.ign.fr/geocontext/mcp
Unrelated, but also take a look at the nice high-density LiDAR point data we have!
https://visionneuse-lidarhd.ign.fr/?px=4441970.281583222&py=...
- delfrrr
Hey, I'm the author of GeoSQL. Found this thread after realizing the repo got 400 stars last week. Ask me anything.