The Cache API: A quick guide

Learn how to use the Cache API to make your application data available offline.

Pete LePage

TheCache APIis a system for storing and retrieving network requests and their corresponding responses. These might be regular requests and responses created in the course of running your application, or they could be created solely for the purpose of storing data for later use.

The Cache API was created to enable service workers to cache network requests so that they can provide fast responses, regardless of network speed or availability. However, the API can also be used as a general storage mechanism.

Where is it available?

The Cache API is available inall modern browsers.It is exposed via the globalcachesproperty, so you can test for the presence of the API with a simple feature detection:

constcacheAvailable='caches'inself;

The Cache API can be accessed from a window, iframe, worker, or service worker.

What can be stored

The caches only store pairs ofRequestand Responseobjects, representing HTTP requests and responses, respectively. However, the requests and responses can contain any kind of data that can be transferred over HTTP.

How much can be stored?

In short,a lot,at least a couple of hundred megabytes, and potentially hundreds of gigabytes or more. Browser implementations vary, but the amount of storage available is usually based on the amount of storage available on the device.

Creating and opening a cache

To open a cache, use thecaches.open(name)method, passing the name of the cache as the single parameter. If the named cache does not exist, it is created. This method returns aPromisethat resolves with theCacheobject.

constcache=awaitcaches.open('my-cache');
// do something with cache...

Adding to a cache

There are three ways to add an item to a cache -add,addAll,andput. All three methods return aPromise.

cache.add

First, there iscache.add().It takes one parameter, either aRequest or a URL (string). It makes a request to the network and stores the response in the cache. If the fetch fails, or if the status code of the response is not in the 200 range, then nothing is stored and thePromiserejects. Note that cross-origin requests not in CORS mode cannot be stored because they return astatusof 0.Such requests can only be stored withput.

// Retreive data.json from the server and store the response.
cache.add(newRequest('/data.json'));

// Retreive data.json from the server and store the response.
cache.add('/data.json');

cache.addAll

Next, there iscache.addAll().It works similarly toadd(),but takes an array ofRequestobjects or URLs (strings). This works similarly to callingcache.addfor each individual request, except that thePromise rejects if any single request is not cached.

consturls=['/weather/today.json', '/weather/tomorrow.json'];
cache.addAll(urls);

In each of these cases, a new entry overwrites any matching existing entry. This uses the same matching rules described in the section on retrieving.

cache.put

Finally, there iscache.put(),which allows you to store either a response from the network, or create and store your ownResponse.It takes two parameters. The first can either be aRequestobject or a URL (string). The second must be aResponse,either from the network, or generated by your code.

// Retrieve data.json from the server and store the response.
cache.put('/data.json');

// Create a new entry for test.json and store the newly created response.
cache.put('/test.json', new Response('{ "foo": "bar" }'));

// Retrieve data.json from the 3rd party site and store the response.
cache.put('https://example /data.json');

Theput()method is more permissive than eitheradd()oraddAll(),and will allow you to store non-CORS responses, or other responses where the status code of the response is not in the 200 range. It will overwrite any previous responses for the same request.

Creating Request objects

Create theRequestobject using a URL for the thing being stored:

constrequest=newRequest('/my-data-store/item-id');

Working with Response objects

TheResponseobject constructor accepts many types of data, including Blobs,ArrayBuffers,FormDataobjects, and strings.

constimageBlob=newBlob([data],{type:'image/jpeg'});
constimageResponse=newResponse(imageBlob);
conststringResponse=newResponse('Helloworld');

You can set the MIME type of aResponseby setting the appropriate header.

constoptions={
headers:{
'Content-Type':'application/json'
}
}
constjsonResponse=newResponse('{}',options);

If you have retrieved aResponseand wish to access its body, there are several helper methods you can use. Each returns aPromisethat resolves with a value of a different type.

For example

constresponse=newResponse('Helloworld');
constbuffer=awaitresponse.arrayBuffer();
console.log(newUint8Array(buffer));
// Uint8Array(11) [72, 101, 108, 108, 111, 32, 119, 111, 114, 108, 100]

Retrieving from a cache

To find an item in a cache, you can use thematchmethod.

constresponse=awaitcache.match(request);
console.log(request,response);

Ifrequestis a string the browser converts it to aRequestby calling new Request(request).The function returns aPromisethat resolves to aResponseif a matching entry is found, orundefinedotherwise.

To determine if twoRequestsmatch, the browser uses more than just the URL. Two requests are considered different if they have different query strings, Varyheaders, or HTTP methods (GET,POST,PUT,etc.).

You can ignore some or all of these things by passing an options object as a second parameter.

constoptions={
ignoreSearch:true,
ignoreMethod:true,
ignoreVary:true
};

constresponse=awaitcache.match(request,options);
// do something with the response

If more than one cached request matches then the one that was created first is returned. If you want to retrieveallmatching responses, you can use cache.matchAll().

constoptions={
ignoreSearch:true,
ignoreMethod:true,
ignoreVary:true
};

constresponses=awaitcache.matchAll(request,options);
console.log(`There are${responses.length}matching responses.`);

As a shortcut you can search over all caches at once by usingcaches.match() instead of callingcache.match()for each cache.

Searching

The Cache API does not provide a way to search for requests or responses except for matching entries against aResponseobject. However, you can implement your own search using filtering or by creating an index.

Filtering

One way to implement your own search is to iterate over all entries and filter down to the ones that you want. Let's say that you want to find all items that have URLs ending with.png.

asyncfunctionfindImages(){
// Get a list of all of the caches for this origin
constcacheNames=awaitcaches.keys();
constresult=[];

for(constnameofcacheNames){
// Open the cache
constcache=awaitcaches.open(name);

// Get a list of entries. Each item is a Request object
for(constrequestofawaitcache.keys()){
// If the request URL matches, add the response to the result
if(request.url.endsWith('.png')){
result.push(awaitcache.match(request));
}
}
}

returnresult;
}

This way you can use any property of theRequestandResponseobjects to filter the entries. Note that this is slow if you search over large sets of data.

Creating an index

The other way to implement your own search is to maintain a separate index of entries that can be searched and store the index in IndexedDB. Since this is the kind of operation that IndexedDB was designed for it has much better performance with large numbers of entries.

If you store the URL of theRequestalongside the searchable properties then you can easily retrieve the correct cache entry after doing the search.

Deleting an item

To delete an item from a cache:

cache.delete(request);

Where request can be aRequestor a URL string. This method also takes the same options object ascache.match,which allows you to delete multiple Request/Responsepairs for the same URL.

cache.delete('/example/file.txt', {ignoreVary: true, ignoreSearch: true});

Deleting a cache

To delete a cache, callcaches.delete(name).This function returns a Promisethat resolves totrueif the cache existed and was deleted, or falseotherwise.

Thanks

Thanks to Mat Scales who wrote the original version of this article, which first appeared on WebFundamentals.