Meet the agents
A team of AI agents — orchestrator, data, spatial, and cartographer — plan, execute, compute, and verify real geospatial work from plain English.
Behind the conversation is a team of specialized AI agents. You describe outcomes in plain English; they coordinate to deliver them.
The roster
- Orchestrator — interprets your intent, plans the steps, and routes work to the right agent.
- Data Agent — finds, fetches, and prepares data: your uploads plus online and MCP catalogs, reprojected and indexed into your workspace.
- Spatial Agent — runs the analysis (vector, raster, remote sensing) on server-side runtimes.
- Cartographer Agent — turns results into presentation-ready maps: choropleths, heatmaps, 3D, and legends.
The agent loop
For each request, the team works in a loop you can follow in the chat panel:
- Plan — the orchestrator interprets your intent and decides the steps.
- Execute & compute — the agents run real geospatial operations on the server-side runtimes.
- Verify — the output is checked and reported back to you.
Because every step is a real, deterministic operation, results are repeatable and auditable — not a black box.
What you can ask for
In plain English, you can have the agents:
- Load data — fetch from open and MCP catalogs, or your uploads.
- Create data — boundaries, points, grids, buffers, and more.
- Analyze data — proximity, watersheds, joins, aggregation, raster and remote-sensing analysis.
- Style the map — choropleths, heatmaps, 3D, and legends.
- Write up results — generate a report or document, or turn a document into a map.
- Write code — Python, SQL, or R, run for you and open to inspect.
You stay in control
- Watch every step as it runs, with clear status and results.
- Provide input when asked — drawing an area, placing a point, or selecting features.
- Inspect and adjust — open any artifact to read or re-run it; check attributes and statistics.
- Refine anytime — each message continues the same conversation about your map.