Discovery scene / Houston, Texas

Discover what
geography knows.

A logistics question becomes a ranked shortlist in minutes. geog.ai screens the region, grades the evidence, and tells you when the answer is only screening-grade.

Where should I build a logistics facility near Houston?

Precomputed from real open map data. Screening-grade evidence, shown plainly.

Houston logistics siting13 evidence categories checked
SCREENING_GRADE
Ranked answer score / 100
Evidence layers
OpenFreeMap Liberty · OpenStreetMap contributors
How geog.ai works

geog.ai investigates
the map for you.

Traditional GIS gives you layers to interpret. geog.ai gives you a decision to interrogate. It makes the assumptions visible, grades the evidence, and tells you when the answer is only screening-grade.

01

Ask in plain language

Describe the decision, the place, and what matters. No layer choreography or specialist syntax. The engine decomposes the question into spatial capabilities.

02

The map is investigated

Fit-for-purpose sources are selected, deterministic spatial tools run, and every result is scored and tagged with its trust level and provenance.

03

Get a ranked, graded answer

Compare candidates by composite score. Open the explanation behind every recommendation. See which evidence categories were available, missing, or commercial-only.

Sample result summary

SCREENING_GRADE · 13 categories checked
Road connectivity91 / 100AUTHORITATIVE
Flood zone exposure84 / 100AUTHORITATIVE
Built environment76 / 100DERIVED
Parcel ownership--COMMERCIAL
What makes it different

Four things you can do
that a map cannot.

01

Rank locations

Every candidate gets a composite score built from deterministic spatial tools — not heuristics or opaque model weights.

02

See the evidence

Road connectivity is observational. Terrain is authoritative. Derived signals are labelled as derived. Each distinction travels with the answer.

03

Ask why

Open any candidate and read the plain-language explanation of why it scored as it did — what helped, what hurt, what was missing.

04

Know what is missing

Gaps are surfaced, not hidden. A missing parcel layer, stale imagery, or a commercial-only source is part of the answer, not a footnote.

Who uses geog.ai
Real estate & site selection Industrial & logistics Insurance & risk Government & public safety Infrastructure investment Environmental compliance
Trust is a feature

Honest about what
the map cannot know.

Every result carries a readiness grade, evidence manifest, and provenance chain. Available, derived, commercial-only, or missing: the status is part of the answer.

Evidence, not confidence theater

Road connectivity is observational. Terrain is authoritative. Built-environment signals are derived. A parcel ownership layer may require a commercial provider. geog.ai keeps those distinctions intact.

Source freshnessEvery evidence source carries a vintage date. The manifest shows when data was collected, not just what it is.
Confidence levelCompleteness and fit are scored separately. A high-confidence source on low-coverage area still yields an honest gap flag.
Why not here?Every eliminated candidate includes a plain-language explanation of which constraints disqualified it and which evidence drove the elimination.
Alternative candidatesThe ranked list always shows the full shortlist, not just the winner. Alternatives with their scores and caveats remain visible.

What the grades mean

DECISION_GRADE supports a decision at scope. WITH_CAVEAT carries a material limitation. SCREENING_GRADE is a shortlist, not a site visit. INSUFFICIENT says stop.

Read the evidence model
No parcel ownership evidence was used in this demonstration. That is a gap, not a footnote.
Commercial evidencePremium data sources — parcel ownership, planning permits, commercial demographics — are quoted separately when a decision requires them. The manifest tells you which sources are commercial-only before you commit.
For people who make places happen

Stop showing stakeholders a map.
Start showing them why.

Site-selection analysts, developers, and operators can move from "what am I looking at?" to "what should we do next?"

Run your first decisionRead the docs
For agents and developers

Your agent already knows how to reason.
Give it geography.

Geog.ai is the spatial intelligence layer your agent is missing. It can UNDERSTAND a question, INVESTIGATE the world, SIMULATE what could happen, OPTIMIZE the decision, and MONITOR the place after it acts.

36 MCP TOOLS43 CAPABILITIES36 SIMULATION ENGINESASYNC JOBS + POLLING
01UNDERSTANDTurn an objective into a spatial plan.
02INVESTIGATEAssemble evidence, context, and constraints.
03SIMULATEModel flood, fire, plume, traffic, heat, or signal.
04OPTIMIZEFind the configuration that survives the constraints.
05MONITORKeep watching for the world to change.
agent / spatial investigationrunning

Find a resilient site for a distribution center near Memphis: low flood exposure, strong terrain, nearby network access, and a viable evacuation route.

get_terrain_profile
Slope and elevation profile · authoritative terrain evidence
2 DC
simulate_flood
100-year inundation scenario · async job queued
job_7f2a
find_nearby
Rail, highway, and critical facilities within radius
1 DC
calculate_evacuation_route
Safe-zone route over active hazard geometry
5 DC
optimize_placement
Ranks candidates against the complete evidence set
20 DC
decision returnedscreening_grade
CANDIDATE RANKED · MEM-03

Recommend Southaven logistics corridor

Best balance of terrain stability, flood resilience, network reach, and emergency access. Two alternatives remain visible with caveats.

87/ 100 composite score
PROVENANCE CHAIN
5 tool results · 4 evidence families · 1 async simulation completed
trust_level: observational + authoritative · freshness attached
Simulation is not a checkbox

Give the agent a world it can test.

Telecom agents can find the minimum tower configuration for 95% coverage. Industrial-safety agents can model a release and identify affected facilities. Emergency planners can simulate evacuation and expose bottlenecks. Drone agents can maximize viewshed; IoT agents can balance coverage, redundancy, and connectivity.

RF coverageFlood SCS-CN + SWMMWildfire cellular automataGaussian / AERMOD / Lagrangian plumesNoise ISO9613HVAC + thermalGrid loadflowEvacuation + trafficCrowd flowOrbital visibilityKriging + tidalWildlife behavior

Decisions & Ranking

Submit an objective. Receive ranked candidates with scored evidence, readiness grades, and plain-language explanations your agent can act on.

Networks & Movement

Shortest routes, evacuation corridors, influence spread, trajectory dwell analysis, and graph queries over spatial networks.

Simulation

Flood, fire spread, plume dispersion, RF coverage, noise, thermal, traffic, and evacuation — 30+ engines accessible as single API calls.

Earth Observation

Satellite change detection, NDVI/NDBI trajectories, spectral analysis, and calibration cycles over any AOI.

Spatial Watches

Register a condition. Get notified when it triggers — flood zone intrusion, development momentum shift, coverage gap emergence.

Evidence & Provenance

Every result carries trust level, freshness, and data lineage. Agents can check evidence adequacy before committing to a conclusion.

01
Agent receives an objective

"Find three candidate sites for a distribution center near Memphis with rail access and low flood risk."

02
Geog.ai builds a spatial plan

Selects evidence sources, acquires terrain, flood zones, transport networks, and parcel data. Runs placement optimization.

03
Returns a graded, explainable answer

Ranked candidates with scores, evidence manifest, readiness grade, and the provenance chain behind every claim.