Simplified HTTP requests
Got is a human-friendly and powerful HTTP request library.
It was created because the popular request
package is bloated:
- Promise & stream API
- Request cancelation
- RFC compliant caching
- Follows redirects
- Retries on failure
- Progress events
- Handles gzip/deflate
- Timeout handling
- Errors with metadata
- JSON mode
- WHATWG URL support
- Electron support
- Instances with custom defaults
- Composable
- Used by ~2000 packages and ~500K repos
- Actively maintained
See how Got compares to other HTTP libraries
$ npm install got
const got = require('got');
(async () => {
try {
const response = await got('sindresorhus.com');
console.log(response.body);
//=> '<!doctype html> ...'
} catch (error) {
console.log(error.response.body);
//=> 'Internal server error ...'
}
})();
const fs = require('fs');
const got = require('got');
got.stream('sindresorhus.com').pipe(fs.createWriteStream('index.html'));
// For POST, PUT, and PATCH methods `got.stream` returns a `stream.Writable`
fs.createReadStream('index.html').pipe(got.stream.post('sindresorhus.com'));
It's a GET
request by default, but can be changed by using different methods or in the options
.
Returns a Promise for a response
object with a body
property, a url
property with the request URL or the final URL after redirects, and a requestUrl
property with the original request URL.
The response object will typically be a Node.js HTTP response stream, however, if returned from the cache it will be a response-like object which behaves in the same way.
The response will also have a fromCache
property set with a boolean value.
Type: string
Object
The URL to request, as a string, a https.request
options object, or a WHATWG URL
.
Properties from options
will override properties in the parsed url
.
If no protocol is specified, it will default to https
.
Type: Object
Any of the https.request
options.
Type: string
Object
When specified, url
will be prepended by baseUrl
.
If you specify an absolute URL, it will skip the baseUrl
.
Very useful when used with got.extend()
to create niche-specific Got instances.
Can be a string or a WHATWG URL
.
Type: Object
Default: {}
Request headers.
Existing headers will be overwritten. Headers set to null
will be omitted.
Type: boolean
Default: false
Returns a Stream
instead of a Promise
. This is equivalent to calling got.stream(url, [options])
.
Type: string
Buffer
stream.Readable
form-data
instance
If you provide this option, got.stream()
will be read-only.
The body that will be sent with a POST
request.
If present in options
and options.method
is not set, options.method
will be set to POST
.
The content-length
header will be automatically set if body
is a string
/ Buffer
/ fs.createReadStream
instance / form-data
instance, and content-length
and transfer-encoding
are not manually set in options.headers
.
Type: string
null
Default: 'utf8'
Encoding to be used on setEncoding
of the response data. If null
, the body is returned as a Buffer
(binary data).
Type: boolean
Default: false
If you provide this option, got.stream()
will be read-only.
If set to true
and Content-Type
header is not set, it will be set to application/x-www-form-urlencoded
.
body
must be a plain object. It will be converted to a query string using (new URLSearchParams(object)).toString()
.
Type: boolean
Default: false
If you use got.stream()
, this option will be ignored.
If set to true
and Content-Type
header is not set, it will be set to application/json
.
Parse response body with JSON.parse
and set accept
header to application/json
. If used in conjunction with the form
option, the body
will the stringified as querystring and the response parsed as JSON.
body
must be a plain object or array and will be stringified.
Type: string
Object
Query string object that will be added to the request URL. This will override the query string in url
.
Type: number
Object
Milliseconds to wait for the server to end the response before aborting the request with got.TimeoutError
error (a.k.a. request
property). By default, there's no timeout.
This also accepts an object
with the following fields to constrain the duration of each phase of the request lifecycle:
lookup
starts when a socket is assigned and ends when the hostname has been resolved. Does not apply when using a Unix domain socket.connect
starts whenlookup
completes (or when the socket is assigned if lookup does not apply to the request) and ends when the socket is connected.secureConnect
starts whenconnect
completes and ends when the handshaking process completes (HTTPS only).socket
starts when the socket is connected. See request.setTimeout.response
starts when the request has been written to the socket and ends when the response headers are received.send
starts when the socket is connected and ends with the request has been written to the socket.request
starts when the request is initiated and ends when the response's end event fires.
Type: number
Object
Default:
- retries:
2
- methods:
GET
PUT
HEAD
DELETE
OPTIONS
TRACE
- statusCodes:
408
413
429
500
502
503
504
- maxRetryAfter:
undefined
An object representing retries
, methods
, statusCodes
and maxRetryAfter
fields for the time until retry, allowed methods, allowed status codes and maximum Retry-After
time.
If maxRetryAfter
is set to undefined
, it will use options.timeout
.
If Retry-After
header is greater than maxRetryAfter
, it will cancel the request.
Delays between retries counts with function 1000 * Math.pow(2, retry) + Math.random() * 100
, where retry
is attempt number (starts from 0).
The retries
property can be a number
or a function
with retry
and error
arguments. The function must return a delay in milliseconds (0
return value cancels retry).
Note: It retries only on the specified methods, status codes, and on these network errors:
ETIMEDOUT
: One of the timeout limits were reached.ECONNRESET
: Connection was forcibly closed by a peer.EADDRINUSE
: Could not bind to any free port.ECONNREFUSED
: Connection was refused by the server.EPIPE
: The remote side of the stream being written has been closed.
Type: boolean
Default: true
Defines if redirect responses should be followed automatically.
Note that if a 303
is sent by the server in response to any request type (POST
, DELETE
, etc.), Got will automatically request the resource pointed to in the location header via GET
. This is in accordance with the spec.
Type: boolean
Default: true
Decompress the response automatically. This will set the accept-encoding
header to gzip, deflate
unless you set it yourself.
If this is disabled, a compressed response is returned as a Buffer
. This may be useful if you want to handle decompression yourself or stream the raw compressed data.
Type: Object
Default: false
Cache adapter instance for storing cached data.
Type: boolean
Default: false
When used in Electron, Got will use electron.net
instead of the Node.js http
module. According to the Electron docs, it should be fully compatible, but it's not entirely. See #443 and #461.
Type: boolean
Default: true
Determines if a got.HTTPError
is thrown for error responses (non-2xx status codes).
If this is disabled, requests that encounter an error status code will be resolved with the response
instead of throwing. This may be useful if you are checking for resource availability and are expecting error responses.
Type: Object<string, Function[]>
Default: {beforeRequest: []}
Hooks allow modifications during the request lifecycle. Hook functions may be async and are run serially.
Type: Function[]
Default: []
Called with the normalized request options. Got will make no further changes to the request before it is sent. This is especially useful in conjunction with got.extend()
and got.create()
when you want to create an API client that, for example, uses HMAC-signing.
See the AWS section for an example.
Note: Modifying the body
is not recommended because the content-length
header has already been computed and assigned.
Note: Progress events, redirect events and request/response events can also be used with promises.
Sets options.stream
to true
.
Returna a duplex stream with additional events:
request
event to get the request object of the request.
Tip: You can use request
event to abort request:
got.stream('github.com')
.on('request', request => setTimeout(() => request.abort(), 50));
The response
event to get the response object of the final request.
The redirect
event to get the response object of a redirect. The second argument is options for the next request to the redirect location.
Progress events for uploading (sending a request) and downloading (receiving a response). The progress
argument is an object like:
{
percent: 0.1,
transferred: 1024,
total: 10240
}
If it's not possible to retrieve the body size (can happen when streaming), total
will be null
.
(async () => {
const response = await got('sindresorhus.com')
.on('downloadProgress', progress => {
// Report download progress
})
.on('uploadProgress', progress => {
// Report upload progress
});
console.log(response);
})();
The error
event emitted in case of a protocol error (like ENOTFOUND
etc.) or status error (4xx or 5xx). The second argument is the body of the server response in case of status error. The third argument is a response object.
Sets options.method
to the method name and makes a request.
Configure a new got
instance with default options
. options
are merged with the parent instance's defaults.options
using got.mergeOptions
.
const client = got.extend({
baseUrl: 'https://example.com',
headers: {
'x-unicorn': 'rainbow'
}
});
client.get('/demo');
/* HTTP Request =>
* GET /demo HTTP/1.1
* Host: example.com
* x-unicorn: rainbow
*/
(async () => {
const client = got.extend({
baseUrl: 'httpbin.org',
headers: {
'x-foo': 'bar'
}
});
const {headers} = (await client.get('/headers', {json: true})).body;
//=> headers['x-foo'] === 'bar'
const jsonClient = client.extend({
json: true,
headers: {
'x-baz': 'qux'
}
});
const {headers: headers2} = (await jsonClient.get('/headers')).body;
//=> headers2['x-foo'] === 'bar'
//=> headers2['x-baz'] === 'qux'
})();
Need more control over the behavior of Got? Check out the got.create()
.
Extends parent options. Avoid using object spread as it doesn't work recursively:
const a = {headers: {cat: 'meow', wolf: ['bark', 'wrrr']}};
const b = {headers: {cow: 'moo', wolf: ['auuu']}};
{...a, ...b} // => {headers: {cow: 'moo', wolf: ['auuu']}}
got.mergeOptions(a, b) // => {headers: {cat: 'meow', cow: 'moo', wolf: ['auuu']}}
Options are deeply merged to a new object. The value of each key is determined as follows:
- If the new property is set to
undefined
, it keeps the old one. - If the parent property is an instance of
URL
and the new value is astring
orURL
, a new URL instance is created:new URL(new, parent)
. - If the new property is a plain
Object
:- If the parent property is a plain
Object
too, both values are merged recursively into a newObject
. - Otherwise, only the new value is deeply cloned.
- If the parent property is a plain
- If the new property is an
Array
, it overwrites the old one with a deep clone of the new property. - Otherwise, the new value is assigned to the key.
Each error contains (if available) statusCode
, statusMessage
, host
, hostname
, method
, path
, protocol
and url
properties to make debugging easier.
In Promise mode, the response
is attached to the error.
When a cache method fails, for example, if the database goes down or there's a filesystem error.
When a request fails. Contains a code
property with error class code, like ECONNREFUSED
.
When reading from response stream fails.
When json
option is enabled, server response code is 2xx, and JSON.parse
fails.
When the server response code is not 2xx. Includes statusCode
, statusMessage
, and redirectUrls
properties.
When the server redirects you more than ten times. Includes a redirectUrls
property, which is an array of the URLs Got was redirected to before giving up.
When given an unsupported protocol.
When the request is aborted with .cancel()
.
When the request is aborted due to a timeout
The promise returned by Got has a .cancel()
method which when called, aborts the request.
(async () => {
const request = got(url, options);
// …
// In another part of the code
if (something) {
request.cancel();
}
// …
try {
await request;
} catch (error) {
if (request.isCanceled) { // Or `error instanceof got.CancelError`
// Handle cancelation
}
// Handle other errors
}
})();
Got implements RFC 7234 compliant HTTP caching which works out of the box in-memory and is easily pluggable with a wide range of storage adapters. Fresh cache entries are served directly from the cache, and stale cache entries are revalidated with If-None-Match
/If-Modified-Since
headers. You can read more about the underlying cache behavior in the cacheable-request
documentation.
You can use the JavaScript Map
type as an in-memory cache:
const got = require('got');
const map = new Map();
(async () => {
let response = await got('sindresorhus.com', {cache: map});
console.log(response.fromCache);
//=> false
response = await got('sindresorhus.com', {cache: map});
console.log(response.fromCache);
//=> true
})();
Got uses Keyv internally to support a wide range of storage adapters. For something more scalable you could use an official Keyv storage adapter:
$ npm install @keyv/redis
const got = require('got');
const KeyvRedis = require('@keyv/redis');
const redis = new KeyvRedis('redis://user:pass@localhost:6379');
got('sindresorhus.com', {cache: redis});
Got supports anything that follows the Map API, so it's easy to write your own storage adapter or use a third-party solution.
For example, the following are all valid storage adapters:
const storageAdapter = new Map();
// Or
const storageAdapter = require('./my-storage-adapter');
// Or
const QuickLRU = require('quick-lru');
const storageAdapter = new QuickLRU({maxSize: 1000});
got('sindresorhus.com', {cache: storageAdapter});
View the Keyv docs for more information on how to use storage adapters.
You can use the tunnel
module with the agent
option to work with proxies:
const got = require('got');
const tunnel = require('tunnel-agent');
got('sindresorhus.com', {
agent: tunnel.httpOverHttp({
proxy: {
host: 'localhost'
}
})
});
If you require different agents for different protocols, you can pass a map of agents to the agent
option. This is necessary because a request to one protocol might redirect to another. In such a scenario, got
will switch over to the right protocol agent for you.
const got = require('got');
const HttpAgent = require('agentkeepalive');
const HttpsAgent = HttpAgent.HttpsAgent;
got('sindresorhus.com', {
agent: {
http: new HttpAgent(),
https: new HttpsAgent()
}
});
You can use the cookie
module to include cookies in a request:
const got = require('got');
const cookie = require('cookie');
got('google.com', {
headers: {
cookie: cookie.serialize('foo', 'bar')
}
});
got('google.com', {
headers: {
cookie: [
cookie.serialize('foo', 'bar'),
cookie.serialize('fizz', 'buzz')
].join(';')
}
});
You can use the form-data
module to create POST request with form data:
const fs = require('fs');
const got = require('got');
const FormData = require('form-data');
const form = new FormData();
form.append('my_file', fs.createReadStream('/foo/bar.jpg'));
got.post('google.com', {
body: form
});
You can use the oauth-1.0a
module to create a signed OAuth request:
const got = require('got');
const crypto = require('crypto');
const OAuth = require('oauth-1.0a');
const oauth = OAuth({
consumer: {
key: process.env.CONSUMER_KEY,
secret: process.env.CONSUMER_SECRET
},
signature_method: 'HMAC-SHA1',
hash_function: (baseString, key) => crypto.createHmac('sha1', key).update(baseString).digest('base64')
});
const token = {
key: process.env.ACCESS_TOKEN,
secret: process.env.ACCESS_TOKEN_SECRET
};
const url = 'https://api.twitter.com/1.1/statuses/home_timeline.json';
got(url, {
headers: oauth.toHeader(oauth.authorize({url, method: 'GET'}, token)),
json: true
});
Requests can also be sent via unix domain sockets. Use the following URL scheme: PROTOCOL://unix:SOCKET:PATH
.
PROTOCOL
-http
orhttps
(optional)SOCKET
- Absolute path to a unix domain socket, for example:/var/run/docker.sock
PATH
- Request path, for example:/v2/keys
got('http://unix:/var/run/docker.sock:/containers/json');
// Or without protocol (HTTP by default)
got('unix:/var/run/docker.sock:/containers/json');
Requests to AWS services need to have their headers signed. This can be accomplished by using the aws4
package. This is an example for querying an "API Gateway" with a signed request.
const AWS = require('aws-sdk');
const aws4 = require('aws4');
const got = require('got');
const credentials = await new AWS.CredentialProviderChain().resolvePromise();
// Create a Got instance to use relative paths and signed requests
const awsClient = got.extend({
baseUrl: 'https://<api-id>.execute-api.<api-region>.amazonaws.com/<stage>/',
hooks: {
beforeRequest: [
async options => {
await credentials.getPromise();
aws4.sign(options, credentials);
}
]
}
});
const response = await awsClient('endpoint/path', {
// Request-specific options
});
You can test your requests by using the nock
module to mock an endpoint:
const got = require('got');
const nock = require('nock');
nock('https://sindresorhus.com')
.get('/')
.reply(200, 'Hello world!');
(async () => {
const response = await got('sindresorhus.com');
console.log(response.body);
//=> 'Hello world!'
})();
If you need real integration tests you can use create-test-server
:
const got = require('got');
const createTestServer = require('create-test-server');
(async () => {
const server = await createTestServer();
server.get('/', 'Hello world!');
const response = await got(server.url);
console.log(response.body);
//=> 'Hello world!'
await server.close();
})();
It's a good idea to set the 'user-agent'
header so the provider can more easily see how their resource is used. By default, it's the URL to this repo. You can omit this header by setting it to null
.
const got = require('got');
const pkg = require('./package.json');
got('sindresorhus.com', {
headers: {
'user-agent': `my-module/${pkg.version} (https://github.com/username/my-module)`
}
});
got('sindresorhus.com', {
headers: {
'user-agent': null
}
});
Bear in mind; if you send an if-modified-since
header and receive a 304 Not Modified
response, the body will be empty. It's your responsibility to cache and retrieve the body contents.
Use got.extend()
to make it nicer to work with REST APIs. Especially if you use the baseUrl
option.
Note: Not to be confused with got.create()
, which has no defaults.
const got = require('got');
const pkg = require('./package.json');
const custom = got.extend({
baseUrl: 'example.com',
json: true,
headers: {
'user-agent': `my-module/${pkg.version} (https://github.com/username/my-module)`
}
});
// Use `custom` exactly how you use `got`
(async () => {
const list = await custom('/v1/users/list');
})();
Need to merge some instances into a single one? Check out got.mergeInstances()
.
- gh-got - Got convenience wrapper to interact with the GitHub API
- gl-got - Got convenience wrapper to interact with the GitLab API
- travis-got - Got convenience wrapper to interact with the Travis API
- graphql-got - Got convenience wrapper to interact with GraphQL
- GotQL - Got convenience wrapper to interact with GraphQL using JSON-parsed queries instead of strings
Sindre Sorhus | Vsevolod Strukchinsky | Alexander Tesfamichael | Luke Childs | Szymon Marczak | Brandon Smith |
MIT