Raster tools
Inspect pixels, clip areas, calculate zonal statistics and work with imagery, elevation, soil and surface products.
Products · GeoRAG, Tools & Skills
Give every agent
a reliable sense of place.
LGM combines evidence retrieval with deterministic geospatial operations, so agents can move from a territorial question to sources, calculations, geometry and a decision-ready answer.
Geospatial intelligence surface
Evidence, computation
and reusable expertise.
GeoRAG supplies grounded context. Tools perform the spatial work. Skills turn proven sequences into repeatable agent capabilities.
Vector tools
Search, filter, intersect, buffer, compare and rank points, lines and polygons with explicit spatial operations.
Geospatial RAG
Retrieve context from catalogs, metadata, indexed sources and prior outputs while preserving provenance for the answer.
Domain Skills
Package tools and decision logic for planning, site selection, risk assessment, monitoring and other repeatable jobs.
Agent workflow
From natural language
to GIS-ready evidence.
A bounded delivery path keeps the work measurable, reviewable and ready to move from validation into production.
- 01
Interpret the task
Turn the user question and territory into a clear analytical objective.
- 02
Retrieve context
Discover relevant catalogs, layers, metadata, documents and statistical indicators.
- 03
Calculate
Use deterministic raster and vector tools for spatial operations instead of asking the model to guess.
- 04
Return evidence
Deliver sources, explanation, structured values, map-ready geometry and usage metadata.
What the agent receives
Results that can be
inspected and reused.
The output is designed to continue into an application or GIS workflow instead of ending as an unsupported paragraph.
- Source and catalog references
- Structured raster and vector results
- Map-ready points, lines and polygons
- Reasoning grounded in retrieved evidence
- Tool-call trace and operation metadata
- Usage attribution for the request and key
Commercial model
Meter the capabilities the agent actually uses.
A partner can expose the complete LGM Skill or selected tools and attribute consumption to the calling application or customer.
- 01Requests to MCP or API tools
- 02Model input and output tokens
- 03Geospatial processing and compute
- 04Data transfer and premium datasets
- 05Wholesale usage or revenue share
Initial Skills
Start with workflows
that repeat.
A Skill should capture a real job, required inputs, tool sequence, quality checks and GIS-ready output—not only a prompt.
City and territorial planning
Combine land use, population, access, constraints and infrastructure into reviewable planning evidence.
Site selection and investment
Rank candidate locations using demand, accessibility, competitors, permits and environmental constraints.
Risk and territory monitoring
Join current imagery, catalogs and spatial rules for repeatable inspection and prioritization.
GeoRAG, Tools & Skills