Place Autocomplete

Select platform: Android iOS JavaScript Web Service

Introduction

Autocomplete is a feature of the Places library in the Maps JavaScript API. You can use autocomplete to give your applications the type-ahead-search behavior of the Google Maps search field. The autocomplete service can match on full words and substrings, resolving place names, addresses, and plus codes. Applications can therefore send queries as the user types, to provide on-the-fly place predictions.

Getting started

Before using the Places library in the Maps JavaScript API, first ensure that the Places API is enabled in the Google Cloud Console, in the same project you set up for the Maps JavaScript API.

To view your list of enabled APIs:

  1. Go to the Google Cloud Console.
  2. Click the Select a project button, then select the same project you set up for the Maps JavaScript API and click Open.
  3. From the list of APIs on the Dashboard, look for Places API.
  4. If you see the API in the list, you’re all set. If the API is not listed, enable it:
    1. At the top of the page, select ENABLE API to display the Library tab. Alternatively, from the left side menu, select Library.
    2. Search for Places API, then select it from the results list.
    3. Select ENABLE. When the process finishes, Places API appears in the list of APIs on the Dashboard.

Loading the library

The Places service is a self-contained library, separate from the main Maps JavaScript API code. To use the functionality contained within this library, you must first load it using the libraries parameter in the Maps API bootstrap URL:

<script async
    src="https://maps.googleapis.com/maps/api/js?key=YOUR_API_KEY&loading=async&libraries=places&callback=initMap">
</script>

See the Libraries Overview for more information.

Summary of classes

The API offers two types of autocomplete widgets, which you can add via the Autocomplete and SearchBox classes respectively. In addition, you can use the AutocompleteService class to retrieve autocomplete results programmatically (see the Maps JavaScript API Reference: AutocompleteService class).

Below is a summary of the classes available:

  • Autocomplete adds a text input field to your web page, and monitors that field for character entries. As the user enters text, autocomplete returns place predictions in the form of a dropdown pick list. When the user selects a place from the list, information about the place is returned to the autocomplete object, and can be retrieved by your application. See the details below.
    An autocomplete text field, and the pick list of place
    predictions supplied as the user enters the search query.
    Figure 1: Autocomplete text field and pick list
    A completed address form.
    Figure 2: Completed address form
  • SearchBox adds a text input field to your web page, in much the same way as Autocomplete. The differences are as follows:
    • The main difference lies in the results that appear in the pick list. SearchBox supplies an extended list of predictions, which can include places (as defined by the Places API) plus suggested search terms. For example, if the user enters 'pizza in new', the pick list may include the phrase 'pizza in New York, NY' as well as the names of various pizza outlets.
    • SearchBox offers fewer options than Autocomplete for restricting the search. In the former, you can bias the search towards a given LatLngBounds. In the latter, you can restrict the search to a particular country and particular place types, as well as setting the bounds. For more information, see below.
    A completed address form.
    Figure 3: A SearchBox presents search terms and place predictions.
    See the details below.
  • You can create an AutocompleteService object to retrieve predictions programmatically. Call getPlacePredictions() to retrieve matching places, or call getQueryPredictions() to retrieve matching places plus suggested search terms. Note: AutocompleteService does not add any UI controls. Instead, the above methods return an array of prediction objects. Each prediction object contains the text of the prediction, as well as reference information and details of how the result matches the user input. See the details below.

Adding an Autocomplete widget

The Autocomplete widget creates a text input field on your web page, supplies predictions of places in a UI pick list, and returns place details in response to a getPlace() request. Each entry in the pick list corresponds to a single place (as defined by the Places API).

The Autocomplete constructor takes two arguments:

  • An HTML input element of type text. This is the input field that the autocomplete service will monitor and attach its results to.
  • An optional AutocompleteOptions argument, which can contain the following properties:
    • An array of data fields to be included in the Place Details response for the user's selected PlaceResult. If the property is not set or if ['ALL'] is passed in, all available fields are returned and billed for (this is not recommended for production deployments). For a list of fields, see PlaceResult.
    • An array of types that specifies an explicit type or a type collection, as listed in the supported types. If no type is specified, all types are returned.
    • bounds is a google.maps.LatLngBounds object specifying the area in which to search for places. The results are biased towards, but not restricted to, places contained within these bounds.
    • strictBounds is a boolean specifying whether the API must return only those places that are strictly within the region defined by the given bounds. The API does not return results outside this region even if they match the user input.
    • componentRestrictions can be used to restrict results to specific groups. Currently, you can use componentRestrictions to filter by up to 5 countries. Countries must be passed as as a two-character, ISO 3166-1 Alpha-2 compatible country code. Multiple countries must be passed as a list of country codes.

      Note: If you receive unexpected results with a country code, verify that you are using a code which includes the countries, dependent territories, and special areas of geographical interest you intend. You can find code information at Wikipedia: List of ISO 3166 country codes or the ISO Online Browsing Platform.

    • placeIdOnly can be used to instruct the Autocomplete widget to retrieve only Place IDs. On calling getPlace() on the Autocomplete object, the PlaceResult made available will only have the place id, types and name properties set. You can use the returned place ID with calls to the Places, Geocoding, Directions or Distance Matrix services.

Constraining Autocomplete predictions

By default, Place Autocomplete presents all place types, biased for predictions near the user's location, and fetches all available data fields for the user's selected place. Set Place Autocomplete options to present more relevant predictions based on your use case.

Set options at construction

The Autocomplete constructor accepts an AutocompleteOptions parameter to set constraints at widget creation. The following example sets the bounds, componentRestrictions, and types options to request establishment type places, favoring those within the specified geographic area and restricting predictions to only places within the United States. Setting the fields option specifies what information to return about the user's selected place.

Call setOptions() to change an option's value for an existing widget.

TypeScript

const center = { lat: 50.064192, lng: -130.605469 };
// Create a bounding box with sides ~10km away from the center point
const defaultBounds = {
  north: center.lat + 0.1,
  south: center.lat - 0.1,
  east: center.lng + 0.1,
  west: center.lng - 0.1,
};
const input = document.getElementById("pac-input") as HTMLInputElement;
const options = {
  bounds: defaultBounds,
  componentRestrictions: { country: "us" },
  fields: ["address_components", "geometry", "icon", "name"],
  strictBounds: false,
};

const autocomplete = new google.maps.places.Autocomplete(input, options);

JavaScript

const center = { lat: 50.064192, lng: -130.605469 };
// Create a bounding box with sides ~10km away from the center point
const defaultBounds = {
  north: center.lat + 0.1,
  south: center.lat - 0.1,
  east: center.lng + 0.1,
  west: center.lng - 0.1,
};
const input = document.getElementById("pac-input");
const options = {
  bounds: defaultBounds,
  componentRestrictions: { country: "us" },
  fields: ["address_components", "geometry", "icon", "name"],
  strictBounds: false,
};
const autocomplete = new google.maps.places.Autocomplete(input, options);

Specify data fields

Specify data fields to avoid being billed for Places Data SKUs you don't need. Include the fields property in the AutocompleteOptions that are passed to the widget constructor, as demonstrated in the previous example, or call setFields() on an existing Autocomplete object.

autocomplete.setFields(["place_id", "geometry", "name"]);

Define biases and search-area boundaries for Autocomplete

You can bias the autocomplete results to favor an approximate location or area, in the following ways:

  • Set the bounds on creation of the Autocomplete object.
  • Change the bounds on an existing Autocomplete.
  • Set the bounds to the map's viewport.
  • Restrict the search to the bounds.
  • Restrict the search to a specific country.

The previous example demonstrates setting bounds at creation. The following examples demonstrate the other biasing techniques.

Change the bounds of an existing Autocomplete

Call setBounds() to change the search area on an existing Autocomplete to rectangular bounds.

TypeScript

const southwest = { lat: 5.6108, lng: 136.589326 };
const northeast = { lat: 61.179287, lng: 2.64325 };
const newBounds = new google.maps.LatLngBounds(southwest, northeast);

autocomplete.setBounds(newBounds);

JavaScript

const southwest = { lat: 5.6108, lng: 136.589326 };
const northeast = { lat: 61.179287, lng: 2.64325 };
const newBounds = new google.maps.LatLngBounds(southwest, northeast);

autocomplete.setBounds(newBounds);
Set the bounds to the map's viewport

Use bindTo() to bias the results to the map's viewport, even while that viewport changes.

TypeScript

autocomplete.bindTo("bounds", map);

JavaScript

autocomplete.bindTo("bounds", map);

Use unbind() to unbind the Autocomplete predictions from the map's viewport.

TypeScript

autocomplete.unbind("bounds");
autocomplete.setBounds({ east: 180, west: -180, north: 90, south: -90 });

JavaScript

autocomplete.unbind("bounds");
autocomplete.setBounds({ east: 180, west: -180, north: 90, south: -90 });

View example

Restrict the search to the current bounds

Set the strictBounds option to restrict the results to the current bounds, whether based on map viewport or rectangular bounds.

autocomplete.setOptions({ strictBounds: true });
Restrict predictions to a specific country

Use the componentRestrictions option or call setComponentRestrictions() to restrict the autocomplete search to a specific set of up to five countries.

TypeScript

autocomplete.setComponentRestrictions({
  country: ["us", "pr", "vi", "gu", "mp"],
});

JavaScript

autocomplete.setComponentRestrictions({
  country: ["us", "pr", "vi", "gu", "mp"],
});

View example

Constrain place types

Use the types option or call setTypes() to constrain predictions to certain place types. This constraint specifies a type or a type collection, as listed in Place Types. If no constraint is specified, all types are returned.

For the value of the types option or the value passed to setTypes(), you can specify either:

  • An array containing up to five values from Table 1 or Table 2 from Place Types. For example:

    types: ['hospital', 'pharmacy', 'bakery', 'country']

    Or:

    autocomplete.setTypes(['hospital', 'pharmacy', 'bakery', 'country']);
  • Any one filter in Table 3 from Place Types. You can only specify a single value from Table 3.

The request will be rejected if:

  • You specify more than five types.
  • You specify any unrecognized types.
  • You mix any types from Table 1 or Table 2 with any filter from Table 3.

The Places Autocomplete demo demonstrates the differences in predictions between different place types.

Visit demo

Getting place information

When a user selects a place from the predictions attached to the autocomplete text field, the service fires a place_changed event. To get place details:

  1. Create an event handler for the place_changed event, and call addListener() on the Autocomplete object to add the handler.
  2. Call Autocomplete.getPlace() on the Autocomplete object, to retrieve a PlaceResult object, which you can then use to get more information about the selected place.

By default, when a user selects a place, autocomplete returns all of the available data fields for the selected place, and you will be billed accordingly. Use Autocomplete.setFields() to specify which place data fields to return. Read more about the PlaceResult object, including a list of place data fields that you can request. To avoid paying for data that you don't need, be sure to use Autocomplete.setFields() to specify only the place data that you will use.

The name property contains the description from Places Autocomplete predictions. You can read more about the description in the Places Autocomplete documentation.

For address forms, it is useful to get the address in structured format. To return the structured address for the selected place, call Autocomplete.setFields() and specify the address_components field.

The following example uses autocomplete to fill the fields in an address form.

TypeScript

function fillInAddress() {
  // Get the place details from the autocomplete object.
  const place = autocomplete.getPlace();
  let address1 = "";
  let postcode = "";

  // Get each component of the address from the place details,
  // and then fill-in the corresponding field on the form.
  // place.address_components are google.maps.GeocoderAddressComponent objects
  // which are documented at http://goo.gle/3l5i5Mr
  for (const component of place.address_components as google.maps.GeocoderAddressComponent[]) {
    // @ts-ignore remove once typings fixed
    const componentType = component.types[0];

    switch (componentType) {
      case "street_number": {
        address1 = `${component.long_name} ${address1}`;
        break;
      }

      case "route": {
        address1 += component.short_name;
        break;
      }

      case "postal_code": {
        postcode = `${component.long_name}${postcode}`;
        break;
      }

      case "postal_code_suffix": {
        postcode = `${postcode}-${component.long_name}`;
        break;
      }

      case "locality":
        (document.querySelector("#locality") as HTMLInputElement).value =
          component.long_name;
        break;

      case "administrative_area_level_1": {
        (document.querySelector("#state") as HTMLInputElement).value =
          component.short_name;
        break;
      }

      case "country":
        (document.querySelector("#country") as HTMLInputElement).value =
          component.long_name;
        break;
    }
  }

  address1Field.value = address1;
  postalField.value = postcode;

  // After filling the form with address components from the Autocomplete
  // prediction, set cursor focus on the second address line to encourage
  // entry of subpremise information such as apartment, unit, or floor number.
  address2Field.focus();
}

JavaScript

function fillInAddress() {
  // Get the place details from the autocomplete object.
  const place = autocomplete.getPlace();
  let address1 = "";
  let postcode = "";

  // Get each component of the address from the place details,
  // and then fill-in the corresponding field on the form.
  // place.address_components are google.maps.GeocoderAddressComponent objects
  // which are documented at http://goo.gle/3l5i5Mr
  for (const component of place.address_components) {
    // @ts-ignore remove once typings fixed
    const componentType = component.types[0];

    switch (componentType) {
      case "street_number": {
        address1 = `${component.long_name} ${address1}`;
        break;
      }

      case "route": {
        address1 += component.short_name;
        break;
      }

      case "postal_code": {
        postcode = `${component.long_name}${postcode}`;
        break;
      }

      case "postal_code_suffix": {
        postcode = `${postcode}-${component.long_name}`;
        break;
      }
      case "locality":
        document.querySelector("#locality").value = component.long_name;
        break;
      case "administrative_area_level_1": {
        document.querySelector("#state").value = component.short_name;
        break;
      }
      case "country":
        document.querySelector("#country").value = component.long_name;
        break;
    }
  }

  address1Field.value = address1;
  postalField.value = postcode;
  // After filling the form with address components from the Autocomplete
  // prediction, set cursor focus on the second address line to encourage
  // entry of subpremise information such as apartment, unit, or floor number.
  address2Field.focus();
}

window.initAutocomplete = initAutocomplete;

View example

Customizing placeholder text

By default, the text field created by the autocomplete service contains standard placeholder text. To modify the text, set the placeholder attribute on the input element:

<input id="searchTextField" type="text" size="50" placeholder="Anything you want!">

Note: The default placeholder text is localized automatically. If you specify your own placeholder value, you must handle the localization of that value in your application. For information on how the Google Maps JavaScript API chooses the language to use, please read the documentation on localization.

See Styling the Autocomplete and SearchBox widgets to customize the widget appearance.

The SearchBox allows users to perform a text-based geographic search, such as 'pizza in New York' or 'shoe stores near robson street'. You can attach the SearchBox to a text field and, as text is entered, the service will return predictions in the form of a drop-down pick list.

SearchBox supplies an extended list of predictions, which can include places (as defined by the Places API) plus suggested search terms. For example, if the user enters 'pizza in new', the pick list may include the phrase 'pizza in New York, NY' as well as the names of various pizza outlets. When a user selects a place from the list, information about that place is returned to the SearchBox object, and can be retrieved by your application.

The SearchBox constructor takes two arguments:

  • An HTML input element of type text. This is the input field that the SearchBox service will monitor and attach its results to.
  • An options argument, which can contain the bounds property: bounds is a google.maps.LatLngBounds object specifying the area in which to search for places. The results are biased towards, but not restricted to, places contained within these bounds.

The following code uses the bounds parameter to bias the results towards places within a particular geographic area, specified via laitude/longitude coordinates.

var defaultBounds = new google.maps.LatLngBounds(
  new google.maps.LatLng(-33.8902, 151.1759),
  new google.maps.LatLng(-33.8474, 151.2631));

var input = document.getElementById('searchTextField');

var searchBox = new google.maps.places.SearchBox(input, {
  bounds: defaultBounds
});

Changing the search area for SearchBox

To change the search area for an existing SearchBox, call setBounds() on the SearchBox object and pass the relevant LatLngBounds object.

View example

Getting place information

When the user selects an item from the predictions attached to the search box, the service fires a places_changed event. You can call getPlaces() on the SearchBox object, to retrieve an array containing several predictions, each of which is a PlaceResult object.

For more information about the PlaceResult object, refer to the documentation on place detail results.

TypeScript

// Listen for the event fired when the user selects a prediction and retrieve
// more details for that place.
searchBox.addListener("places_changed", () => {
  const places = searchBox.getPlaces();

  if (places.length == 0) {
    return;
  }

  // Clear out the old markers.
  markers.forEach((marker) => {
    marker.setMap(null);
  });
  markers = [];

  // For each place, get the icon, name and location.
  const bounds = new google.maps.LatLngBounds();

  places.forEach((place) => {
    if (!place.geometry || !place.geometry.location) {
      console.log("Returned place contains no geometry");
      return;
    }

    const icon = {
      url: place.icon as string,
      size: new google.maps.Size(71, 71),
      origin: new google.maps.Point(0, 0),
      anchor: new google.maps.Point(17, 34),
      scaledSize: new google.maps.Size(25, 25),
    };

    // Create a marker for each place.
    markers.push(
      new google.maps.Marker({
        map,
        icon,
        title: place.name,
        position: place.geometry.location,
      })
    );

    if (place.geometry.viewport) {
      // Only geocodes have viewport.
      bounds.union(place.geometry.viewport);
    } else {
      bounds.extend(place.geometry.location);
    }
  });
  map.fitBounds(bounds);
});

JavaScript

// Listen for the event fired when the user selects a prediction and retrieve
// more details for that place.
searchBox.addListener("places_changed", () => {
  const places = searchBox.getPlaces();

  if (places.length == 0) {
    return;
  }

  // Clear out the old markers.
  markers.forEach((marker) => {
    marker.setMap(null);
  });
  markers = [];

  // For each place, get the icon, name and location.
  const bounds = new google.maps.LatLngBounds();

  places.forEach((place) => {
    if (!place.geometry || !place.geometry.location) {
      console.log("Returned place contains no geometry");
      return;
    }

    const icon = {
      url: place.icon,
      size: new google.maps.Size(71, 71),
      origin: new google.maps.Point(0, 0),
      anchor: new google.maps.Point(17, 34),
      scaledSize: new google.maps.Size(25, 25),
    };

    // Create a marker for each place.
    markers.push(
      new google.maps.Marker({
        map,
        icon,
        title: place.name,
        position: place.geometry.location,
      }),
    );
    if (place.geometry.viewport) {
      // Only geocodes have viewport.
      bounds.union(place.geometry.viewport);
    } else {
      bounds.extend(place.geometry.location);
    }
  });
  map.fitBounds(bounds);
});

View example

See Styling the Autocomplete and SearchBox widgets to customize the widget appearance.

Programmatically retrieving Place Autocomplete Service predictions

To retrieve predictions programmatically, use the AutocompleteService class. AutocompleteService does not add any UI controls. Instead, it returns an array of prediction objects, each containing the text of the prediction, reference information, and details of how the result matches the user input. This is useful if you want more control over the user interface than is offered by the Autocomplete and SearchBox described above.

AutocompleteService exposes the following methods:

  • getPlacePredictions() returns place predictions. Note: A 'place' can be an establishment, geographic location, or prominent point of interest, as defined by the Places API.
  • getQueryPredictions() returns an extended list of predictions, which can include places (as defined by the Places API) plus suggested search terms. For example, if the user enters 'pizza in new', the pick list may include the phrase 'pizza in New York, NY' as well as the names of various pizza outlets.

Both of the above methods return an array of prediction objects of the following form:

  • description is the matched prediction.
  • distance_meters is the distance in meters of the place from the specified AutocompletionRequest.origin.
  • matched_substrings contains a set of substrings in the description that match elements in the user's input. This is useful for highlighting those substrings in your application. In many cases, the query will appear as a substring of the description field.
    • length is the length of the substring.
    • offset is the character offset, measured from the beginning of the description string, at which the matched substring appears.
  • place_id is a textual identifier that uniquely identifies a place. To retrieve information about the place, pass this identifier in the placeId field of a Place Details request. Learn more about how to reference a place with a place ID.
  • terms is an array containing elements of the query. For a place, each element will typically make up a portion of the address.
    • offset is the character offset, measured from the beginning of the description string, at which the matched substring appears.
    • value is the matching term.

The example below executes a query prediction request for the phrase 'pizza near' and displays the result in a list.

TypeScript

// This example retrieves autocomplete predictions programmatically from the
// autocomplete service, and displays them as an HTML list.
// This example requires the Places library. Include the libraries=places
// parameter when you first load the API. For example:
// <script src="https://maps.googleapis.com/maps/api/js?key=YOUR_API_KEY&libraries=places">
function initService(): void {
  const displaySuggestions = function (
    predictions: google.maps.places.QueryAutocompletePrediction[] | null,
    status: google.maps.places.PlacesServiceStatus
  ) {
    if (status != google.maps.places.PlacesServiceStatus.OK || !predictions) {
      alert(status);
      return;
    }

    predictions.forEach((prediction) => {
      const li = document.createElement("li");

      li.appendChild(document.createTextNode(prediction.description));
      (document.getElementById("results") as HTMLUListElement).appendChild(li);
    });
  };

  const service = new google.maps.places.AutocompleteService();

  service.getQueryPredictions({ input: "pizza near Syd" }, displaySuggestions);
}

declare global {
  interface Window {
    initService: () => void;
  }
}
window.initService = initService;

JavaScript

// This example retrieves autocomplete predictions programmatically from the
// autocomplete service, and displays them as an HTML list.
// This example requires the Places library. Include the libraries=places
// parameter when you first load the API. For example:
// <script src="https://maps.googleapis.com/maps/api/js?key=YOUR_API_KEY&libraries=places">
function initService() {
  const displaySuggestions = function (predictions, status) {
    if (status != google.maps.places.PlacesServiceStatus.OK || !predictions) {
      alert(status);
      return;
    }

    predictions.forEach((prediction) => {
      const li = document.createElement("li");

      li.appendChild(document.createTextNode(prediction.description));
      document.getElementById("results").appendChild(li);
    });
  };

  const service = new google.maps.places.AutocompleteService();

  service.getQueryPredictions({ input: "pizza near Syd" }, displaySuggestions);
}

window.initService = initService;

CSS

HTML

<html>
  <head>
    <title>Retrieving Autocomplete Predictions</title>
    <script src="https://polyfill.io/v3/polyfill.min.js?features=default"></script>

    <link rel="stylesheet" type="text/css" href="./style.css" />
    <script type="module" src="./index.js"></script>
  </head>
  <body>
    <p>Query suggestions for 'pizza near Syd':</p>
    <ul id="results"></ul>
    <!-- Replace Powered By Google image src with self hosted image. https://developers.google.com/maps/documentation/places/web-service/policies#other_attribution_requirements -->
    <img
      class="powered-by-google"
      src="https://storage.googleapis.com/geo-devrel-public-buckets/powered_by_google_on_white.png"
      alt="Powered by Google"
    />

    <!-- 
      The `defer` attribute causes the callback to execute after the full HTML
      document has been parsed. For non-blocking uses, avoiding race conditions,
      and consistent behavior across browsers, consider loading using Promises.
      See https://developers.google.com/maps/documentation/javascript/load-maps-js-api
      for more information.
      -->
    <script
      src="https://maps.googleapis.com/maps/api/js?key=AIzaSyB41DRUbKWJHPxaFjMAwdrzWzbVKartNGg&callback=initService&libraries=places&v=weekly"
      defer
    ></script>
  </body>
</html>

Try Sample

View example

Session tokens

AutocompleteService.getPlacePredictions() can use session tokens (if implemented) to group together autocomplete requests for billing purposes. Session tokens group the query and selection phases of a user autocomplete search into a discrete session for billing purposes. The session begins when the user starts typing a query, and concludes when they select a place. Each session can have multiple queries, followed by one place selection. Once a session has concluded, the token is no longer valid. Your app must generate a fresh token for each session. We recommend using session tokens for all autocomplete sessions. If the sessionToken parameter is omitted, or if you reuse a session token, the session is charged as if no session token was provided (each request is billed separately).

You can use the same session token to make a single Place Details request on the place that results from a call to AutocompleteService.getPlacePredictions(). In this case, the autocomplete request is combined with the Place Details request, and the call is charged as a regular Place Details request. There is no charge for the autocomplete request.

Be sure to pass a unique session token for each new session. Using the same token for more than one Autocomplete session will invalidate those Autocomplete sessions, and all Autocomplete request in the invalid sessions will be charged individually using Autocomplete Per Request SKU. Read more about session tokens.

The following example shows creating a session token, then passing it in an AutocompleteService (the displaySuggestions() function has been omitted for brevity):

// Create a new session token.
var sessionToken = new google.maps.places.AutocompleteSessionToken();

// Pass the token to the autocomplete service.
var autocompleteService = new google.maps.places.AutocompleteService();
autocompleteService.getPlacePredictions({
  input: 'pizza near Syd',
  sessionToken: sessionToken
},
displaySuggestions);

Be sure to pass a unique session token for each new session. Using the same token for more than one session will result in each request being billed individually.

Read more about session tokens.

Styling the Autocomplete and SearchBox widgets

By default, the UI elements provided by Autocomplete and SearchBox are styled for inclusion on a Google map. You may want to adjust the styling to suit your own site. The following CSS classes are available. All classes listed below apply to both the Autocomplete and the SearchBox widgets.

A graphical illustration of the CSS classes for the Autocomplete and
      SearchBox widgets
CSS classes for Autocomplete and SearchBox widgets
CSS class Description
pac-container The visual element containing the list of predictions returned by the Place Autocomplete service. This list appears as a dropdown list below the Autocomplete or SearchBox widget.
pac-icon The icon displayed to the left of each item in the list of predictions.
pac-item An item in the list of predictions supplied by the Autocomplete or SearchBox widget.
pac-item:hover The item when the user hovers their mouse pointer over it.
pac-item-selected The item when the user selects it via the keyboard. Note: Selected items will be a member of this class and of the pac-item class.
pac-item-query A span inside a pac-item that is the main part of the prediction. For geographic locations, this contains a place name, like 'Sydney', or a street name and number, like '10 King Street'. For text-based searches such as 'pizza in New York', it contains the full text of the query. By default, the pac-item-query is colored black. If there is any additional text in the pac-item, it is outside pac-item-query and inherits its styling from pac-item. It is colored gray by default. The additional text is typically an address.
pac-matched The part of the returned prediction that matches the user’s input. By default, this matched text is highlighted in bold text. Note that the matched text may be anywhere within pac-item. It is not necessarily part of pac-item-query, and it could be partly within pac-item-query as well as partly in the remaining text in pac-item.

Use the Place Picker component

Note: This sample uses an open source library. See the README for support and feedback related to the library.

Try web components. Use the Place Picker web component to enable text input that allows end users to search for a specific address or place using autocomplete.

GIF with search box. User starts to type an address as input and a dropdown with related
      addresses appears. User clicks an address from the dropdown and the search box fills
      in the rest of the address.
Figure 1: Text input to search for a specific address or place using autocomplete

Place Autocomplete optimization

This section describes best practices to help you make the most of the Place Autocomplete service.

Here are some general guidelines:

  • The quickest way to develop a working user interface is to use the Maps JavaScript API Autocomplete widget, Places SDK for Android Autocomplete widget, or Places SDK for iOS Autocomplete UI control
  • Develop an understanding of essential Place Autocomplete data fields from the start.
  • Location biasing and location restriction fields are optional but can have a significant impact on autocomplete performance.
  • Use error handling to make sure your app degrades gracefully if the API returns an error.
  • Make sure your app handles when there is no selection and offers users a way to continue.

Cost optimization best practices

Basic cost optimization

To optimize the cost of using the Place Autocomplete service, use field masks in Place Details and Place Autocomplete widgets to return only the place data fields you need.

Advanced cost optimization

Consider programmatic implementation of Place Autocomplete in order to access Per Request pricing and request Geocoding API results about the selected place instead of Place Details. Per Request pricing paired with Geocoding API is more cost-effective than Per Session (session-based) pricing if both of the following conditions are met:

  • If you only need the latitude/longitude or address of the user's selected place, the Geocoding API delivers this information for less than a Place Details call.
  • If users select an autocomplete prediction within an average of four Autocomplete predictions requests or fewer, Per Request pricing may be more cost-effective than Per Session pricing.
For help selecting the Place Autocomplete implementation that fits your needs, select the tab that corresponds to your answer to the following question.

Does your application require any information other than the address and latitude/longitude of the selected prediction?

Yes, needs more details

Use session-based Place Autocomplete with Place Details.
Since your application requires Place Details such as the place name, business status, or opening hours, your implementation of Place Autocomplete should use a session token (programmatically or built into the JavaScript, Android, or iOS widgets) for a total cost of $0.017 per session plus applicable Places Data SKUs depending on which place data fields you request.1

Widget implementation
Session management is automatically built into the JavaScript, Android, or iOS widgets. This includes both the Place Autocomplete requests and the Place Details request on the selected prediction. Be sure to specify the fields parameter in order to ensure you are only requesting the place data fields you need.

Programmatic implementation
Use a session token with your Place Autocomplete requests. When requesting Place Details about the selected prediction, include the following parameters:

  1. The place ID from the Place Autocomplete response
  2. The session token used in the Place Autocomplete request
  3. The fields parameter specifying the place data fields you need

No, needs only address and location

Geocoding API could be a more cost-effective option than Place Details for your application, depending on the performance of your Place Autocomplete usage. Every application's Autocomplete efficiency varies depending on what users are entering, where the application is being used, and whether performance optimization best practices have been implemented.

In order to answer the following question, analyze how many characters a user types on average before selecting a Place Autocomplete prediction in your application.

Do your users select a Place Autocomplete prediction in four or fewer requests, on average?

Yes

Implement Place Autocomplete programmatically without session tokens and call Geocoding API on the selected place prediction.
Geocoding API delivers addresses and latitude/longitude coordinates for $0.005 per request. Making four Place Autocomplete - Per Request requests costs $0.01132 so the total cost of four requests plus a Geocoding API call about the selected place prediction would be $0.01632 which is less than the Per Session Autocomplete price of $0.017 per session.1

Consider employing performance best practices to help your users get the prediction they're looking for in even fewer characters.

No

Use session-based Place Autocomplete with Place Details.
Since the average number of requests you expect to make before a user selects a Place Autocomplete prediction exceeds the cost of Per Session pricing, your implementation of Place Autocomplete should use a session token for both the Place Autocomplete requests and the associated Place Details request for a total cost of $0.017 per session.1

Widget implementation
Session management is automatically built into the JavaScript, Android, or iOS widgets. This includes both the Place Autocomplete requests and the Place Details request on the selected prediction. Be sure to specify the fields parameter in order to ensure you are only requesting Basic Data fields.

Programmatic implementation
Use a session token with your Place Autocomplete requests. When requesting Place Details about the selected prediction, include the following parameters:

  1. The place ID from the Place Autocomplete response
  2. The session token used in the Place Autocomplete request
  3. The fields parameter specifying Basic Data fields such as address and geometry

Consider delaying Place Autocomplete requests
You can employ strategies such as delaying a Place Autocomplete request until the user has typed in the first three or four characters so that your application makes fewer requests. For example, making Place Autocomplete requests for each character after the user has typed the third character means that if the user types seven characters then selects a prediction for which you make one Geocoding API request, the total cost would be $0.01632 (4 * $0.00283 Autocomplete Per Request + $0.005 Geocoding).1

If delaying requests can get your average programmatic request below four, you can follow the guidance for performant Place Autocomplete with Geocoding API implementation. Note that delaying requests can be perceived as latency by the user who might be expecting to see predictions with every new keystroke.

Consider employing performance best practices to help your users get the prediction they're looking for in fewer characters.


  1. Costs listed here are in USD. Please refer to the Google Maps Platform Billing page for full pricing information.

Performance best practices

The following guidelines describe ways to optimize Place Autocomplete performance:

  • Add country restrictions, location biasing, and (for programmatic implementations) language preference to your Place Autocomplete implementation. Language preference is not needed with widgets since they pick language preferences from the user's browser or mobile device.
  • If Place Autocomplete is accompanied by a map, you can bias location by map viewport.
  • In situations when a user does not choose one of the Autocomplete predictions, generally because none of those predictions are the desired result-address, you can re-use the original user input to attempt to get more relevant results:
    • If you expect the user to enter only address information, re-use the original user input in a call to the Geocoding API.
    • If you expect the user to enter queries for a specific place by name or address, use a Find Place request. If results are only expected in a specific region, use location biasing.
    Other scenarios when it's best to fall back to the Geocoding API include:
    • Users inputting subpremise addresses in countries where Place Autocomplete support of subpremise addresses is incomplete, e.g. Czechia, Estonia and Lithuania. For example, the Czech address "Stroupežnického 3191/17, Praha" yields a partial prediction in Place Autocomplete.
    • Users inputting addresses with road-segment prefixes like "23-30 29th St, Queens" in New York City or "47-380 Kamehameha Hwy, Kaneohe" on the island of Kauai in Hawai'i.

Usage limits and policies

Quotas

For quota and pricing information, see the Usage and Billing documentation for the Places API.

Policies

Use of the Places Library, Maps JavaScript API must be in accordance with the policies described for the Places API.