Programmable
Geography
EmbedEarth is a programmable layer for the physical world. Search and compute across 1M+ geographic features and categories with AI, maps, places, routing, and open infrastructure.
$ npm i @embedearth/cliA global, queryable base map built from open sources. Use it as-is or make it your own.
Find places, addresses, and areas with one familiar API.
Understand demand, coverage, and how people interact with the physical world.
A drop-in Places API for finding and understanding real-world locations.
Run spatial jobs, tiling, and models on managed infrastructure—scale from a laptop to a cluster without leaving your data behind.
Directions, distance matrices, and travel times without proprietary map infrastructure.
Translate addresses, coordinates, and landmarks with one familiar API.
Bring Planet, feature tiles, routes, and spatial compute into one MapLibre surface.
Explore street-level imagery and see the world from the ground.
AI SEARCH
Search an area by what it looks like.
EmbedEarth AI Search uses AI to find areas with a similar visual character. Search satellite or street-level imagery directly, without needing a perfectly structured database first.
Explore AI Search →BUILD WITH E2
SDK, CLI, and MCP,
open and on npm
Everything ships as open-source packages. Install one, inspect the code, and start building programmable geography in minutes.
SDK
JavaScript and TypeScript client for search, routing, maps, and spatial compute.
@embedearth/sdkView on npm →CLI
Search, route, and compute from the terminal with one command tree.
@embedearth/cliView on npm →MCP
Model Context Protocol server that gives AI agents live geo tools.
@embedearth/mcpView on npm →Switch without rewriting
your map experience
import { E2Map, useSearch } from "@e2/react";
export function CoffeeMap() {
const { results } = useSearch({ q: "coffee near me" });
return <E2Map apiKey={import.meta.env.VITE_E2_KEY} markers={results} />;
}Open source from day one
EmbedEarth is built in the open so your map stack stays transparent, inspectable, and yours to control. Read the code, contribute, self-host, or use EmbedEarth Cloud. No black boxes. No lock-in.
◉ View on GitHubOPEN-SOURCE ALTERNATIVE
The open-source alternative to
Google Maps APIs
EmbedEarth is a drop-in alternative to Google Maps APIs—Places, Autocomplete, Geocoding, and Routes—built on open data and infrastructure you control. Swap the endpoint, keep your code, and never pay a lock-in tax again.
Drop-in APIs
Familiar Places, Autocomplete, and Routes interfaces—migrate from Google without rewriting your app.
Open data
EmbedEarth is built on open places, addresses, and regions—transparent, portable, and yours.
Infrastructure you control
Self-host EmbedEarth or use our cloud. Your keys, your data, your terms.
AI-POWERED GEOGRAPHY
Geography your AI can
actually use
EmbedEarth pairs 1M+ geographic features with AI-native interfaces—search areas by what they look like, give agents live geo tools, and compute over the physical world in one call.
AI Search
Find areas with a similar visual character from satellite or street-level imagery.
Explore AI Search →MCP server
Expose EmbedEarth search, routing, and compute as tools your AI agents can call.
View MCP docs →Trends & Compute
Measure demand, coverage, and change across every EmbedEarth feature and region.
Explore Trends →EMBEDEARTH CLOUD
All your geo needs,
in one cloud
Get every geo primitive from EmbedEarth Cloud—search, tiles, routes, storage, and compute—behind one API key. Start locally, scale to production, and stay on infrastructure you control.
Search & Places
Places, Autocomplete, and Geocoding APIs.
Tiles & Maps
Vector tiles and MapLibre-ready surfaces.
Routes
Directions, matrices, and travel times.
Storage & Compute
Datasets, snapshots, and spatial jobs.