This plugin makes it easy to create a JSON API for your entries (and other element types) in Craft CMS.
It’s powered by Phil Sturgeon’s excellent Fractal package.
This plugin requires Craft CMS 4.0 or later.
You can install this plugin from the Plugin Store or with Composer.
Go to the Plugin Store in your project’s Control Panel and search for “Element API”. Then click on the “Install” button in its modal window.
Open your terminal and run the following commands:
# go to the project directory
cd /path/to/my-project.test
# tell Composer to load the plugin
composer require craftcms/element-api
# tell Craft to install the plugin
./craft plugin/install element-api
To define your API endpoints, create a new element-api.php
file within your config/
folder. This file should return an array with an endpoints
key, which defines your site’s API endpoints.
Within the endpoints
array, keys are URL patterns, and values are functions that define the endpoint configurations.
<?php
use craft\elements\Entry;
use craft\helpers\UrlHelper;
return [
'endpoints' => [
'news.json' => function() {
return [
'elementType' => Entry::class,
'criteria' => ['section' => 'news'],
'transformer' => function(Entry $entry) {
return [
'title' => $entry->title,
'url' => $entry->url,
'jsonUrl' => UrlHelper::url("news/{$entry->id}.json"),
'summary' => $entry->summary,
];
},
];
},
'news/<entryId:\d+>.json' => function($entryId) {
return [
'elementType' => Entry::class,
'criteria' => ['id' => $entryId],
'one' => true,
'transformer' => function(Entry $entry) {
return [
'title' => $entry->title,
'url' => $entry->url,
'summary' => $entry->summary,
'body' => $entry->body,
];
},
];
},
]
];
Endpoint configuration arrays can contain the following settings:
The class name of the Fractal resource that should be used to serve the request. If this isn’t set, it will default to craft\elementapi\resources\ElementResource
. (All of the following configuration settings are specific to that default class.)
The class name of the element type that the API should be associated with. Craft’s built-in element type classes are:
craft\elements\Asset
craft\elements\Category
craft\elements\Entry
craft\elements\GlobalSet
craft\elements\MatrixBlock
craft\elements\Tag
craft\elements\User
'elementType' => craft\elements\Entry::class,
An array of parameters that should be set on the Element Query that will be fetching the elements.
'criteria' => [
'section' => 'news',
'type' => 'article',
],
The content type the endpoint responses should have.
'contentType' => 'application/foo+json',
By default, the content type will be:
application/javascript
for endpoints that define a JSONP callbackapplication/feed+json
for endpoints where the serializer is set tojsonFeed
application/json
for everything else
The transformer that should be used to define the data that should be returned for each element. If you don’t set this, the default transformer will be used, which includes all of the element’s direct attribute values, but no custom field values.
// Can be set to a function
'transformer' => function(craft\elements\Entry $entry) {
return [
'title' => $entry->title,
'id' => $entry->id,
'url' => $entry->url,
];
},
// Or a string/array that defines a Transformer class configuration
'transformer' => 'MyTransformerClassName',
// Or a Transformer class instance
'transformer' => new MyTransformerClassName(),
Your custom transformer class would look something like this:
<?php
use craft\elements\Entry;
use League\Fractal\TransformerAbstract;
class MyTransformerClassName extends TransformerAbstract
{
public function transform(Entry $entry)
{
return [
// ...
];
}
}
Whether only the first matching element should be returned. This is set to false
by default, meaning that all matching elements will be returned.
'one' => true,
Whether the results should be paginated. This is set to true
by default, meaning that only a subset of the matched elements will be included in each response, accompanied by additional metadata that describes pagination information.
'paginate' => false,
The max number of elements that should be included in each page, if pagination is enabled. By default this is set to 100.
'elementsPerPage' => 10,
The query string param name that should be used to identify which page is being requested. By default this is set to 'page'
.
'pageParam' => 'pg',
Note that it cannot be set to 'p'
because that’s the parameter Craft uses to check the requested path.
The key that the elements should be nested under in the response data. By default this will be 'data'
.
'resourceKey' => 'entries',
Any custom meta values that should be included in the response data.
'meta' => [
'description' => 'Recent news from Happy Lager',
],
The serializer that should be used to format the returned data.
Possible values are:
'array'
(default) – formats data using the ArraySerializer.'dataArray'
– formats data using the DataArraySerializer.'jsonApi'
– formats data using the JsonApiSerializer.'jsonFeed'
– formats data based on JSON Feed V1.1 (see the JSON Feed example below).- A custom serializer instance.
The include names that should be included for the current request, if any.
'includes' => (array)Craft::$app->request->getQueryParam('include'),
Note that this setting requires a custom transformer class that’s prepped to handle includes:
class MyTransformerClassName extends TransformerAbstract
{
protected $availableIncludes = ['author'];
public function includeAuthor(Entry $entry)
{
return $this->item($entry->author, function(User $author) {
return [
'id' => $author->id,
'name' => $author->name,
];
});
}
// ...
}
The include names that should be excluded for the current request, which would otherwise have been included (e.g. if they were listed as a default include), if any.
'excludes' => 'author',
Like includes
, this setting requires a custom transformer class.
If this is set, a JSONP response will be returned with an application/javascript
content type, using this setting value as the callback function.
For example, if you set this to:
'callback' => 'foo',
Then the response will look like:
foo({ /* ... */ });
Note that if you set this, the jsonOptions
and pretty
settings will be ignored.
The value of the $options
argument that will be passed to json_encode()
when preparing the response. By default JSON_UNESCAPED_SLASHES | JSON_UNESCAPED_UNICODE
will be passed.
'jsonOptions' => JSON_UNESCAPED_UNICODE,
Shortcut for adding JSON_PRETTY_PRINT
to jsonOptions
.
'pretty' => true,
Whether the output should be cached, and for how long.
Possible values are:
true
(default) – results are cached for the duration specified by thecacheDuration
Craft config setting, or until a relevant element is saved or deleted.false
– results are never cached.- an integer – results are cached for the given number of seconds.
- a interval spec string – results are cached for the duration specified.
Note that the onBeforeSendData
event does not get triggered when the cache is warm.
'cache' => 'PT1M', // one minute
The key that responses should be cached with, if something custom is needed.
URL patterns can contain dynamic subpatterns in the format of <subpatternName:regex>
, where subpatternName
is the name of the subpattern, and regex
is a valid regular expression. For example, the URL pattern “news/<entryId:\d+>.json
” will match URLs like news/100.json
. You can also use the tokens {handle}
and {slug}
within your regular expression, which will be replaced with the appropriate regex patterns for matching handles and element slugs.
Any subpattern matches in the URL pattern will be mapped to the endpoint config function’s arguments. For example, if a URL pattern contains an entryId
subpattern, then you can add an $entryId
argument to your endpoint config function, and whatever matches the URL subpattern will be passed to $entryId
.
'news/<entryId:\d+>.json' => function($entryId) {
return [
'elementType' => craft\elements\Entry::class,
'criteria' => ['id' => $entryId],
'one' => true,
];
},
You can specify default values for your endpoint configuration settings by adding a defaults
key alongside your endpoints
key (not within it).
use craft\elements\Entry;
return [
'defaults' => [
'elementType' => Entry::class,
'elementsPerPage' => 10,
'pageParam' => 'pg',
'transformer' => function(Entry $entry) {
return [
'title' => $entry->title,
'id' => $entry->id,
'url' => $entry->url,
];
},
],
'endpoints' => [
'news.json' => function() {
return [
'criteria' => ['section' => 'news'],
]
},
'news/<entryId:\d+>.json' => function($entryId) {
return [
'criteria' => ['id' => $entryId],
'one' => true,
];
},
]
];
Here are a few endpoint examples, and what their response would look like.
'ingredients.json' => function() {
return [
'criteria' => ['section' => 'ingredients'],
'elementsPerPage' => 10,
'transformer' => function(craft\elements\Entry $entry) {
return [
'title' => $entry->title,
'url' => $entry->url,
'jsonUrl' => UrlHelper::url("ingredients/{$entry->slug}.json"),
];
},
'pretty' => true,
];
},
{
"data": [
{
"title": "Gin",
"url": "/ingredients/gin",
"jsonUrl": "/ingredients/gin.json"
},
{
"title": "Tonic Water",
"url": "/ingredients/tonic-water",
"jsonUrl": "/ingredients/tonic-water.json"
},
// ...
],
"meta": {
"pagination": {
"total": 66,
"count": 10,
"per_page": 10,
"current_page": 1,
"total_pages": 7,
"links": {
"next": "/ingredients.json?p=2"
}
}
}
}
'ingredients/<slug:{slug}>.json' => function($slug) {
return [
'criteria' => [
'section' => 'ingredients',
'slug' => $slug
],
'one' => true,
'transformer' => function(craft\elements\Entry $entry) {
// Create an array of all the photo URLs
$photos = [];
foreach ($entry->photos->all() as $photo) {
$photos[] = $photo->url;
}
return [
'title' => $entry->title,
'url' => $entry->url,
'description' => (string)$entry->description,
'photos' => $photos
];
},
'pretty' => true,
];
},
{
"title": "Gin",
"url": "/ingredients/gin",
"description": "<p>Gin is a spirit which derives its predominant flavour from juniper berries.</p>",
"photos": [
"/images/drinks/GinAndTonic1.jpg"
]
}
Here’s how to set up a JSON Feed (Version 1.1) for your site with Element API.
Note that photos
, body
, summary
, and tags
are imaginary custom fields.
'feed.json' => function() {
return [
'serializer' => 'jsonFeed',
'elementType' => craft\elements\Entry::class,
'criteria' => ['section' => 'news'],
'transformer' => function(craft\elements\Entry $entry) {
$image = $entry->photos->one();
return [
'id' => (string)$entry->id,
'url' => $entry->url,
'title' => $entry->title,
'content_html' => (string)$entry->body,
'summary' => $entry->summary,
'image' => $image ? $image->url : null,
'date_published' => $entry->postDate->format(\DateTime::ATOM),
'date_modified' => $entry->dateUpdated->format(\DateTime::ATOM),
'authors' => [
['name' => $entry->author->name],
],
'language' => $entry->getSite()->language,
'tags' => array_map('strval', $entry->tags->all()),
];
},
'meta' => [
'description' => 'Recent news from Happy Lager',
],
'pretty' => true,
];
},
{
"version": "https://jsonfeed.org/version/1",
"title": "Happy Lager",
"home_page_url": "http://domain.com/",
"feed_url": "http://domain.com/feed.json",
"description": "Craft demo site",
"items": [
{
"id": "24",
"url": "http://domain.com/news/the-future-of-augmented-reality",
"title": "The Future of Augmented Reality",
"content_html": "<p>Nam libero tempore, cum soluta nobis est eligendi ...</p>",
"date_published": "2016-05-07T00:00:00+00:00",
"date_modified": "2016-06-03T17:43:36+00:00",
"author": {
"name": "Liz Murphy"
},
"tags": [
"augmented reality",
"futurism"
]
},
{
"id": "4",
"url": "http://domain.com/news/barrel-aged-digital-natives",
"title": "Barrel Aged Digital Natives",
"content_html": "<p>Nam libero tempore, cum soluta nobis est eligendi ...</p>",,
"date_published": "2016-05-06T00:00:00+00:00",
"date_modified": "2017-05-18T13:20:27+00:00",
"author": {
"name": "Liz Murphy"
},
"tags": [
"barrel-aged"
]
},
// ...
]
}