Address-to-coordinate resolution that expects landmark-relative addresses, returns a confidence score on every match, and lets you cache what you resolve.
Geocoding turns a written address into a latitude and longitude that software can plot, measure, and route between. Reverse geocoding runs the other way, taking a coordinate from a device GPS or a dropped pin and returning the nearest addressable place.
Almost every location feature rests on one of these two operations: showing a delivery on a map, calculating distance to the nearest store, or validating that an address a customer typed actually exists.
A large share of Indian addresses are landmark-relative rather than street-numbered. An address reading "opposite Ganesh temple, behind the bus depot" carries no house number and no formal street name, yet it is completely unambiguous to a local delivery rider. Building numbering is frequently non-sequential, the same locality name recurs across cities, and a PIN code can cover an area containing thousands of premises.
Geocoders built around Western street grids handle this by falling back to the centroid of whatever area they matched. The failure is quiet: the API returns a valid-looking coordinate with no indication it is a guess.
maps.guru returns a confidence score and match type on every result, so your code can tell a premise-level match from an area centroid and prompt for a map pin instead of silently accepting the drift.
Use search when you have a complete address string and want it resolved in one call, which is the right shape for cleaning up an existing database of customer records. Use autocomplete in any field where a person is typing: it is optimised for keystroke latency and returns ranked partial matches, so a checkout form resolves an address in a few characters rather than the full string.
Use reverse geocoding when you start from a coordinate, typically a device GPS fix or a pin the user dragged on a map, and need the nearest addressable place.
Resolve a written Indian address to coordinates, and handle the landmark-relative case where a precise premise match does not exist.
Create a free account. A default key with the geocoding scope is generated automatically, and the free tier includes 10,000 geocoding requests a month with no card.
Send the full address string to the search endpoint with your key. Include the PIN code when you have it: it is the single strongest disambiguator for Indian addresses, because locality names repeat across cities.
Each result carries a confidence score and a match type. A premise or street match is safe to use directly. A locality-level match means the geocoder fell back to an area centroid.
On a locality-level match, prompt the user to confirm the pin on a map rather than accepting the centroid. For delivery routing, fall back to PIN-code zoning instead of treating the centroid as a doorstep.
Store the resolved coordinate against your address record. Repeat deliveries to the same address then cost nothing, which is where most of the savings in an address-heavy application come from.
The vocabulary behind address resolution, in plain language.
Geocoding is the conversion of a written address or place name into a latitude and longitude coordinate pair. Reverse geocoding is the inverse: it takes a coordinate and returns the nearest address or place. A geocoding API exposes both operations over HTTP so an application can resolve user-entered addresses to points it can plot, measure, or route between.
Indian addresses are frequently landmark-relative rather than street-numbered. A delivery address may read "opposite Ganesh temple, behind the bus depot" with no house number and no formal street name. Building numbering is often non-sequential, the same locality name recurs across cities, and PIN codes cover areas large enough to contain thousands of premises. Geocoders trained mainly on Western street grids resolve these to the centroid of a locality, which can be a kilometre or more from the real premises.
Accuracy varies by address completeness rather than by city tier. A full address with premise, street, locality, and PIN code usually resolves to the correct street segment. A landmark-only address typically resolves to the locality centroid. Every maps.guru geocoding response carries a confidence score and a match type so you can decide programmatically whether to accept the coordinate, prompt the user to drop a pin, or fall back to PIN-code-level delivery zoning.
Search resolves a complete address string in one call and is the right choice for batch processing existing records. Autocomplete is optimised for keystroke latency and returns ranked partial matches as a user types, which is what an address field in a checkout flow needs. Reverse geocoding takes a coordinate, usually from a device GPS or a dropped pin, and returns the nearest addressable place.
The free tier includes 10,000 geocoding requests per month with no credit card. The Pro tier raises that to 200,000 per month. Requests beyond the quota are declined rather than billed as overage, so a traffic spike cannot produce an unexpected invoice.
Yes. maps.guru permits caching resolved coordinates in your own database indefinitely, which materially reduces cost for address books and repeat deliveries. This differs from some incumbent providers whose terms restrict how long geocoded coordinates may be stored outside their platform. Check the terms of any provider you compare against before you design a caching layer.
Join developers worldwide who trust maps.guru for their mapping needs. Start free, no credit card required.