Geocoding API ServicesOct 20, 2026By Toolverly Editorial

Geocoding API Services in 2026: Google Maps Geocoding API vs Mapbox Geocoding vs HERE Geocoding & Search vs TomTom Geocoding API vs OpenCage

Compare five geocoding APIs by address-match evidence, coordinate meaning, storage options and review requirements for business address intake.

Person reviewing website analytics on a laptop at a desk.

For service-area address intake, choose a geocoding API by the evidence needed to accept a location. A returned coordinate is not proof of a deliverable house, a matched apartment or a position inside your service area. Confidence fields from different APIs do not share a universal scale. Preserve the provider’s match type and component evidence before applying your own acceptance policy.

Compare the information behind an address match

API Evidence examined Decision implication
Google Maps Geocoding API types; geometry.location_type; partial_match Treat geometry precision and address match separately
Mapbox Geocoding properties.match_code; properties.coordinates.accuracy Keep component match evidence alongside the selected point
HERE Geocoding & Search resultType; scoring.queryScore/fieldScore; access Separate query match, house type and entrance position
TomTom Geocoding API type; score; matchConfidence.score in Maps v1 Do not conflate relevance score and match confidence
OpenCage components._type; geometry; confidence Do not treat confidence as a universal correctness probability

Google Maps Geocoding API

Google’s v3 response guide distinguishes result types, geometry location types and partial_match. A rooftop-style geometry classification addresses a different question from whether all requested address components matched. Keep both dimensions available to the reviewer and do not accept a street result merely because it has valid coordinates.

Read the storage, display and attribution policies and usage/billing guide for the intended workflow. A persistent customer-address database, map display and per-request quota need separate review. Storage permissions and the applicable SKU should be confirmed for the intended use.

Mapbox Geocoding

The v6 documentation exposes address match_code, coordinate accuracy and routable-point information where available. Component-level evidence can make a review screen more useful than a single point. Do not silently replace an address point with a routable point without recording which meaning the downstream workflow expects.

The same documentation distinguishes temporary from permanent use through a request option. Pricing separates Geocoding from Search Box and other Mapbox products. Confirm the authorized storage mode before comparing costs for saved locations. Optional result fields are not guaranteed for every address.

HERE Geocoding & Search

HERE’s getting-started guide includes typed results and access coordinates. Its scoring explanation separates query/field matches; extra irrelevant query text can lower a query score even when the address result is useful. This is a counterexample to treating one threshold as an address correctness probability.

Examine the plan limits and excluded-use-case page for your workflow. The retrieved Base Plan pricing page did not expose a usable numeric rate table in our check, so no numeric HERE quote is inferred. Confirm the relevant plan and use-case restrictions for the deployment.

TomTom Geocoding API

TomTom’s current TomTom Maps v1 Geocode reference documents typed results and both a relevance score and match-confidence structure. Pin the API family and version when building adapters: the separately documented Orbis v2 API must not be assumed to have an identical response contract.

The current pricing destination uses API-specific usage estimation. Compare the endpoint you will actually call and retain type and component details. Treat relevance and match-confidence scores as different fields.

OpenCage

OpenCage’s API documentation explains that confidence reflects precision and is not the result ranking or the way to determine matched place type. A large, correctly identified city can have a low precision score. Use the components and their type to distinguish the city result from a house-level result; do not turn that score directly into a percentage of address correctness.

The pricing page distinguishes recurring and request-volume arrangements with rate/usage limits. Check whether your workload is a one-off import or ongoing intake. Choose a request-volume arrangement and rate limit that match that intake pattern.

Choose an explainable review path

For a component-level address-review screen, examine Mapbox’s v6 match codes and HERE’s field scores while preserving their distinct meanings. Google’s v3 result/geometry/partial-match signals support a different explicit review record. Pin TomTom’s API family before implementing its relevance and match-confidence fields. OpenCage is a candidate when its components and precision evidence meet the workflow, but its confidence must not be interpreted as a correctness percentage. These differences describe the information available to a review workflow; they do not establish a regional accuracy ranking.

An address-review process should distinguish incomplete house numbers, ambiguous cities, apartments and service-area borders, while retaining the provider’s original match evidence. Geocoding feeds location selection; it does not replace a fleet platform or the operational scheduling in field-service software.

Prepared by Toolverly Editorial with AI-assisted research. Product details were checked against the linked official sources on October 20, 2026.

Photo by Myriam Jessier on Unsplash. Read about our publication or send a correction.