Infrastructure for Physical World AI Agents.
Mireye turns a latitude and longitude — or a US street address — into sourced, structured data about terrain, land cover, buildings, roads, water, and risk. Three HTTP endpoints (/v1/ask, /v1/fetch, /v1/geocode) and an MCP server, with a citation attached to every field.
What Mireye provides
250+ named fields across 7 layers, plus 15 presets — terrain and soils, flood and wildfire risk, land cover, buildings and roads, the electric grid and gas network, water supply, solar and wind resource, natural hazards, parcels and boundaries. Every value carries a source name, source URL, dataset vintage, fetched-at timestamp, and confidence rating.
Coverage: United States only — latitude 18 to 72, longitude -180 to -65 (CONUS, Alaska, Hawaii, and US territories). Data sources are federal: USGS, NOAA, USDA, USFS, USFWS, FEMA, EPA, EIA, NREL, LBNL, US Census, BLM, USACE, and more.
For AI agents
Discover the full catalog with one public, no-auth call: GET https://api.mireye.com/v1/meta/fields. Ask natural-language questions at https://api.mireye.com/v1/ask or fetch named fields at https://api.mireye.com/v1/fetch (bearer token) — both accept a US street address instead of a coordinate, or resolve one first at https://api.mireye.com/v1/geocode. Connect an agent client to the hosted MCP server at https://api.mireye.com/mcp.
FAQ
What does Mireye actually do?
Mireye turns a latitude and longitude — or a US street address — into sourced, structured data about terrain, land cover, buildings, water, and risk. Three HTTP endpoints (/v1/ask, /v1/fetch, /v1/geocode) and an MCP server, with a citation attached to every field.
What does it cost?
Free for now. We are in early access while we tune coverage and rate limits with design partners. Pricing will be announced when we move out of early access.
What region does Mireye cover?
United States only at launch. Most of the catalog is US federal (USGS, FEMA, NOAA, USDA, BLM, EPA, EIA and more), with a few open and licensed additions such as Overture and Sentinel-2. Every value carries its own source, so you can always see which one you are relying on.
Where does the data come from?
Federal datasets, primarily. Every field in every response returns a source name, a source URL, a fetched_at timestamp, and a confidence level so the agent can pass provenance straight through to the end user or auditor.
How fresh is the data?
It varies by layer. Elevation (USGS 3DEP) is effectively static. FEMA flood maps refresh on FEMA's cadence. USDA cropland updates yearly. Every response includes the fetched_at timestamp so the agent always knows what it is looking at.
Does Mireye train on my queries?
No. Your queries and the answers we return are not used to train any model.
Documentation · Agent guide (skills.md) · Compare /ask vs a frontier model