MCP SERVERTurn addresses into coordinates, routes and travel-time comparisons, then hand back a rendered map image — including multi-stop routes reordered across up to 25 intermediate stops.
LABSCO SUMMARY
The plain endpoints are all here, but the composite tools are the reason to install it: route planning geocodes the stops, optimises the visit order and returns each leg in one call, and the comparison tool searches, pulls details and measures distance in the same pass.
A server over Google's mapping APIs: geocoding, place search and details, directions, distance matrix, elevation, timezone, weather, air quality, static map images and a local-rank grid scan. Fourteen of the tools are single calls; four compose several into one.
A Google Maps API key in GOOGLE_MAPS_API_KEY, and Node.js. Enable Places API (New) and Routes API in the Google Cloud console first — the place and routing tools call those specifically, and a key without them fails on the tools most people install this for. Billing follows Google's own per-API pricing, so the composite tools that fan out to several endpoints cost accordingly.
One command plus a key — npx @cablate/mcp-google-map --stdio, then supply credentials
USE CASES
Hand over the addresses and get them geocoded, reordered for the shortest route, and returned leg by leg.
Compare them side by side — ratings, details and distance from where you are — instead of opening each one.
Explore the area across several place types at once, or search by type and radius when you know what you want.
Geocode up to fifty in one call, then reverse-geocode any coordinates that came from somewhere else.
Check the forecast, the air quality and the local time for the destination before committing to a date.
Generate the map image with the markers and path drawn on it, returned inline in the conversation.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
maps_air_qualityGet air quality for a location — AQI index, pollutant concentrations, and health recommendations by demographic group (elderly, children, athletes, pregnant women, etc.). Use when the user asks 'is the air safe', 'should I wear a mask', 'good for outdoor exercise', or is planning travel for someone with respiratory/heart conditions. Coverage: global including Japan (unlike weather). Returns both universal AQI and local index (EPA for US, AEROS for Japan, etc.).
maps_batch_geocodeGeocode multiple addresses in one call — up to 50 addresses, returns coordinates for each. Use when the user provides a list of addresses and needs all their coordinates, e.g. 'geocode these 10 offices' or 'get coordinates for all these restaurants'. For more than 50, use the CLI batch-geocode command instead.
maps_compare_placesCompare multiple places side-by-side in one call — searches by query, gets details for each result, and optionally calculates distance from your location. Use when the user asks 'which restaurant should I pick', 'compare these hotels', or needs a decision table. Replaces the manual chain of search-places → place-details → distance-matrix.
maps_directionsGet step-by-step navigation directions between two points with route details. Use when the user asks 'how do I get from A to B?' and needs the route summary, total distance, estimated travel time, or turn-by-turn instructions. Supports departure/arrival times and multiple travel modes.
maps_distance_matrixCalculate travel distances and durations between multiple origins and destinations in a single request. Use for comparing travel options — e.g., 'which hotel is closest to the office?' or batch distance calculations. Supports driving, walking, bicycling, and transit modes.
maps_elevationGet elevation (meters above sea level) for geographic coordinates. Use when the user asks 'how high is this place', 'is this area flood-prone', or needs altitude for hiking/cycling route profiles. Also useful for real estate risk assessment — low elevation near water suggests flood risk.
maps_explore_areaExplore what's around a location in one call — searches multiple place types, gets details for the top results, and returns a categorized summary. Use when the user asks 'what's around here', 'explore the area near my hotel', or needs a quick overview of a neighborhood. Replaces the manual chain of geocode → search-nearby → place-details. For trip planning: use search_places first to get geographically spread anchor points, then call this tool around each anchor (e.g. 'Gion, Kyoto') — never pass just the city name, as it clusters all results in one area. After results, call static_map to visualize.
maps_geocodeConvert an address, city name, or landmark into GPS coordinates (latitude/longitude). Use when you need coordinates for a location described in text — for example, to provide a center point for search_nearby or a starting point for maps_directions.
maps_local_rank_trackerTrack a business's local search ranking across a geographic grid (like LocalFalcon). Searches the same keyword(s) from multiple coordinates around a center point to see how rank varies by location. Supports up to 3 keywords for batch scanning. Returns rank at each grid point, top-3 competitors per point, and summary metrics (ARP, ATRP, SoLV). Useful for local SEO analysis.
maps_place_detailsGet comprehensive details for a specific place using its Google Maps place_id. Use after search_nearby or maps_search_places to get full information including reviews, phone number, website, and opening hours. Set maxPhotos (1-10) to include photo URLs — omit or set to 0 for no photos (saves tokens).
maps_plan_routePlan an optimized multi-stop route in one call — geocodes all stops, uses Routes API waypoint optimization (up to 25 intermediate stops) to find the most efficient visit order, and returns directions for each leg. Use when the user says 'visit these 5 places efficiently', 'plan a route through A, B, C', or needs a multi-stop itinerary. Replaces the manual chain of geocode → distance-matrix → directions. For multi-day trips: create one plan_route call per day with stops that follow a geographic arc (e.g. east→west) rather than mixing distant areas. After results, call static_map to visualize the route.
maps_reverse_geocodeConvert GPS coordinates (latitude/longitude) into a human-readable street address. Use when you have coordinates from another tool's output or a user's shared location and need the actual address.
maps_search_along_routeSearch for places along a route between two points — restaurants, cafes, gas stations, etc. ranked by minimal detour time. Use for trip planning to find meals, rest stops, or attractions between landmarks without backtracking. Internally computes the route, then searches along it. Essential for building itineraries where stops should feel 'on the way' rather than 'detour to'.
maps_search_nearbyFind places near a specific location by type (e.g., restaurants, cafes, hotels). Use when the user wants to discover what's around a given address or coordinates, such as 'find coffee shops near Times Square' or 'what hotels are near the airport'. Supports filtering by place type, search radius, minimum rating, and whether currently open.
maps_search_placesSearch for places using a free-text query like 'sushi restaurants in Tokyo' or 'best coffee shops near Central Park'. More flexible than search_nearby — supports natural language queries, optional location bias, rating filters, and open-now filtering. Use when the user describes what they're looking for in words rather than by type and coordinates.
maps_static_mapGenerate a map image with markers, paths, or routes — returned as an inline image the user can see directly in chat. PROACTIVELY call this tool after explore_area, plan_route, search_nearby, or directions to visualize results on a map — don't wait for the user to ask. Use markers from search results and path from route data. Supports roadmap, satellite, terrain, and hybrid views. Max 640x640 pixels.
maps_timezoneGet the timezone and current local time for a location. Use when the user asks 'what time is it in Tokyo', needs to coordinate a meeting across timezones, or is planning travel across timezone boundaries. Returns timezone ID, UTC/DST offsets, and computed local time.
maps_weatherGet weather for a location — current conditions, daily forecast (10 days), or hourly forecast (240 hours). Use when the user asks 'what's the weather in Paris', is planning outdoor activities, or needs to pack for a trip. Coverage: most regions supported, but China, Japan, South Korea, Cuba, Iran, North Korea, Syria are unavailable.
BEFORE YOU INSTALL
18 tools when we connected. GOOGLE_MAPS_ENABLED_TOOLS can register a subset to save context; leave it unset or set it to * for all 18. maps_place_photo is on the roadmap, not yet available, and every tool is read-only.
CONTEXT COST
Rated by the size of the 18 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
npx @cablate/mcp-google-map --stdio{
"mcpServers": {
"google-maps": {
"command": "npx",
"args": [
"-y",
"@cablate/mcp-google-map",
"--stdio"
],
"env": {
"GOOGLE_MAPS_API_KEY": "<YOUR_GOOGLE_MAPS_API_KEY>"
}
}
}
}Keep this terminal running. Streamable HTTP is the default transport when --stdio is left off.
Start the server in HTTP mode
npx @cablate/mcp-google-map --port 3000 --apikey "<YOUR_APIKEY>"Point your MCP client at it
{
"mcpServers": {
"google-maps": {
"type": "http",
"url": "http://localhost:3000/mcp"
}
}
}Keep this terminal running. The key can stay off the command line here: a key sent in the X-Google-Maps-API-Key header takes priority over one given to the server.
Start the server in HTTP mode
npx @cablate/mcp-google-map --port 3000Send your key from the client in a header
{
"mcpServers": {
"mcp-google-map": {
"type": "http",
"url": "http://localhost:3000/mcp",
"headers": {
"X-Google-Maps-API-Key": "<YOUR_GOOGLE_MAPS_API_KEY>"
}
}
}
}npx @cablate/mcp-google-map exec geocode '{"address":"Tokyo Tower"}'MCP servers
david-pivonka
baidu-maps
iamvibhorsingh
jagan-shanmugam