The needle npm package is a minimalistic HTTP client for Node.js, designed for simplicity and ease of use. It provides a straightforward interface for making HTTP requests and handling responses, supporting both callbacks and streams. It is lightweight and has built-in support for multipart form-data, which makes it suitable for file uploads and other form-related tasks.
Making HTTP GET requests
This feature allows you to perform HTTP GET requests to retrieve data from a specified URL. The callback function receives an error object and the response object, which includes the status code and the response body.
const needle = require('needle');
needle.get('https://api.example.com', (error, response) => {
if (!error && response.statusCode == 200)
console.log(response.body);
});Making HTTP POST requests
This feature is used to send HTTP POST requests with data to a specified URL. The data can be an object, which needle will automatically encode as application/x-www-form-urlencoded or multipart/form-data, depending on the content.
const needle = require('needle');
const data = { foo: 'bar' };
needle.post('https://api.example.com', data, (error, response) => {
if (!error && response.statusCode == 200)
console.log(response.body);
});Streaming support
Needle supports streaming, which allows you to pipe the response stream to another stream, such as a file write stream. This is useful for handling large amounts of data or streaming file downloads.
const needle = require('needle');
const fs = require('fs');
const stream = needle.get('https://api.example.com/stream');
stream.pipe(fs.createWriteStream('output.txt'));Handling multipart form-data
Needle can handle multipart form-data, which is often used for file uploads. The data object can include file buffers, streams, or paths, and needle will handle the multipart encoding for you.
const needle = require('needle');
const data = {
file: { file: 'path/to/file.jpg', content_type: 'image/jpeg' },
other_field: 'value'
};
needle.post('https://api.example.com/upload', data, { multipart: true }, (error, response) => {
if (!error && response.statusCode == 200)
console.log('Upload successful');
});Axios is a promise-based HTTP client for the browser and Node.js. It provides a rich set of features like interceptors, automatic JSON data transformation, and cancellation. Compared to needle, axios is more feature-rich and has a larger footprint.
Request is a simplified HTTP request client that offers a wide range of features including OAuth signing, form uploads, and cookies. It is more complex and heavier than needle but has been deprecated as of 2020.
Got is a human-friendly and powerful HTTP request library for Node.js. It supports promises and async/await, streams, retries, and more. Got is more comprehensive than needle and is designed to be a more modern alternative with a focus on simplicity and composability.
Node-fetch is a light-weight module that brings the Fetch API to Node.js. It is a minimal and straightforward API that closely mirrors the browser's fetch API. Compared to needle, node-fetch is more aligned with web standards.
Superagent is a small progressive client-side HTTP request library, and Node.js module with the same API, sporting many high-level HTTP client features. It is more expressive than needle and includes features like a fluent API, chaining, and more.
The leanest and most handsome HTTP client in the Nodelands.
var needle = require('needle');
needle.get('http://www.google.com', function(error, response) {
if (!error && response.statusCode == 200)
console.log(response.body);
});
Callbacks not floating your boat? Needle got your back.
var data = {
file: '/home/johnlennon/walrus.png',
content_type: 'image/png'
};
// the callback is optional, and needle returns a `readableStream` object
// that triggers a 'done' event when the request/response process is complete.
needle
.post('https://my.server.com/foo', data, { multipart: true })
.on('readable', function() { /* eat your chunks */ })
.on('done', function(err) {
console.log('Ready-o!');
})
From version 2.0.x up, Promises are also supported. Just call needle() directly and you'll get a native Promise object.
needle('put', 'https://hacking.the.gibson/login', { password: 'god' }, { json: true })
.then(function(response) {
return doSomethingWith(response)
})
.catch(function(err) {
console.log('Call the locksmith!')
})
With only two real dependencies, Needle supports:
iconv-liteAnd yes, Mr. Wayne, it does come in black.
This makes Needle an ideal alternative for performing quick HTTP requests in Node, either for API interaction, downloading or uploading streams of data, and so on. If you need OAuth, AWS support or anything fancier, you should check out mikeal's request module.
$ npm install needle
// using promises
needle('get', 'https://server.com/posts/123')
.then(function(resp) {
// ...
})
.catch(function(err) {
// ...
});
// with callback
needle.get('ifconfig.me/all.json', function(error, response, body) {
if (error) throw error;
// body is an alias for `response.body`,
// that in this case holds a JSON-decoded object.
console.log(body.ip_addr);
});
// no callback, using streams
needle.get('https://google.com/images/logo.png')
.pipe(fs.createWriteStream('logo.png'))
.on('done', function(err) {
console.log('Pipe finished!');
});
As you can see, you can use Needle with Promises or without them. When using Promises or when a callback is passed, the response's body will be buffered and written to response.body, and the callback will be fired when all of the data has been collected and processed (e.g. decompressed, decoded and/or parsed).
When no callback is passed, however, the buffering logic will be skipped but the response stream will still go through Needle's processing pipeline, so you get all the benefits of post-processing while keeping the streamishness we all love from Node.
Depending on the response's Content-Type, Needle will either attempt to parse JSON or XML streams, or, if a text response was received, will ensure that the final encoding you get is UTF-8.
You can also request a gzip/deflated/brotli response, which, if sent by the server, will be processed before parsing or decoding is performed. (Note: brotli is only supported on Node 10.16.0 or above, and will not be requested or processed on earlier versions.)
needle.get('http://stackoverflow.com/feeds', { compressed: true }, function(err, resp) {
console.log(resp.body); // this little guy won't be a Gzipped binary blob
// but a nice object containing all the latest entries
});
Or in anti-callback mode, using a few other options:
var options = {
compressed : true, // sets 'Accept-Encoding' to 'gzip, deflate, br'
follow_max : 5, // follow up to five redirects
rejectUnauthorized : true // verify SSL certificate
}
var stream = needle.get('https://backend.server.com/everything.html', options);
// read the chunks from the 'readable' event, so the stream gets consumed.
stream.on('readable', function() {
while (data = this.read()) {
console.log(data.toString());
}
})
stream.on('done', function(err) {
// if our request had an error, our 'done' event will tell us.
if (!err) console.log('Great success!');
})
(> 2.0.x)Calling needle() directly returns a Promise. Besides method and url, all parameters are optional, although when sending a post, put or patch request you will get an error if data is not present.
needle('get', 'http://some.url.com')
.then(function(resp) { console.log(resp.body) })
.catch(function(err) { console.error(err) })
Except from the above, all of Needle's request methods return a Readable stream, and both options and callback are optional. If passed, the callback will return three arguments: error, response and body, which is basically an alias for response.body.
needle.head('https://my.backend.server.com', {
open_timeout: 5000 // if we're not able to open a connection in 5 seconds, boom.
}, function(err, resp) {
if (err)
console.log('Shoot! Something is wrong: ' + err.message)
else
console.log('Yup, still alive.')
})
needle.get('google.com/search?q=syd+barrett', function(err, resp) {
// if no http:// is found, Needle will automagically prepend it.
});
var options = {
headers: { 'X-Custom-Header': 'Bumbaway atuna' }
}
needle.post('https://my.app.com/endpoint', 'foo=bar', options, function(err, resp) {
// you can pass params as a string or as an object.
});
var nested = {
params: {
are: {
also: 'supported'
}
}
}
needle.put('https://api.app.com/v2', nested, function(err, resp) {
console.log('Got ' + resp.bytes + ' bytes.') // another nice treat from this handsome fella.
});
Same behaviour as PUT.
var options = {
username: 'fidelio',
password: 'x'
}
needle.delete('https://api.app.com/messages/123', null, options, function(err, resp) {
// in this case, data may be null, but you need to explicity pass it.
});
Generic request. This not only allows for flexibility, but also lets you perform a GET request with data, in which case will be appended to the request as a query string, unless you pass a json: true option (read below).
var params = {
q : 'a very smart query',
page : 2
}
needle.request('get', 'forum.com/search', params, function(err, resp) {
if (!err && resp.statusCode == 200)
console.log(resp.body); // here you go, mister.
});
Now, if you set pass json: true among the options, Needle won't set your params as a querystring but instead send a JSON representation of your data through the request's body, as well as set the Content-Type and Accept headers to application/json.
needle.request('get', 'forum.com/search', params, { json: true }, function(err, resp) {
if (resp.statusCode == 200) console.log('It worked!');
});
The Readable stream object returned by the above request methods emits the following events, in addition to the regular ones (e.g. end, close, data, pipe, readable).
'response'response <http.IncomingMessage>Emitted when the underlying http.ClientRequest emits a response event. This is after the connection is established and the header received, but before any of it is processed (e.g. authorization required or redirect to be followed). No data has been consumed at this point.
'redirect'location <String>Indicates that the a redirect is being followed. This means that the response code was a redirect (301, 302, 303, 307) and the given redirect options allowed following the URL received in the Location header.
'header'statusCode <Integer>headers <Object>Triggered after the header has been processed, and just before the data is to be consumed. This implies that no redirect was followed and/or authentication header was received. In other words, we got a "valid" response.
'done' (previously 'end')exception <Error> (optional)Emitted when the request/response process has finished, either because all data was consumed or an error ocurred somewhere in between. Unlike a regular stream's end event, Needle's done will be fired either on success or on failure, which is why the first argument may be an Error object. In other words:
var resp = needle.get('something.worthy/of/being/streamed/by/needle');
resp.pipe(someWritableStream);
resp.on('done', function(err) {
if (err) console.log('An error ocurred: ' + err.message);
else console.log('Great success!');
})
'err'exception <Error>Emitted when an error ocurrs. This should only happen once in the lifecycle of a Needle request.
'timeout'type <String>Emitted when an timeout error occurs. Type can be either 'open', 'response', or 'read'. This will called right before aborting the request, which will also trigger an err event, a described above, with an ECONNRESET (Socket hang up) exception.
For information about options that've changed, there's always the changelog.
agent : Uses an http.Agent of your choice, instead of the global, default one. Useful for tweaking the behaviour at the connection level, such as when doing tunneling (see below for an example).json : When true, sets content type to application/json and sends request body as JSON string, instead of a query string.open_timeout: (or timeout) Returns error if connection takes longer than X milisecs to establish. Defaults to 10000 (10 secs). 0 means no timeout.response_timeout: Returns error if no response headers are received in X milisecs, counting from when the connection is opened. Defaults to 0 (no response timeout).read_timeout: Returns error if data transfer takes longer than X milisecs, once response headers are received. Defaults to 0 (no timeout).follow_max : (or follow) Number of redirects to follow. Defaults to 0. See below for more redirect options.multipart : Enables multipart/form-data encoding. Defaults to false. Use it when uploading files.proxy : Forwards request through HTTP(s) proxy. Eg. proxy: 'http://user:pass@proxy.server.com:3128'. For more advanced proxying/tunneling use a custom agent, as described below.headers : Object containing custom HTTP headers for request. Overrides defaults described below.auth : Determines what to do with provided username/password. Options are auto, digest or basic (default). auto will detect the type of authentication depending on the response headers.stream_length: When sending streams, this lets you manually set the Content-Length header --if the stream's bytecount is known beforehand--, preventing ECONNRESET (socket hang up) errors on some servers that misbehave when receiving payloads of unknown size. Set it to 0 and Needle will get and set the stream's length for you, or leave unset for the default behaviour, which is no Content-Length header for stream payloads.localAddress: , IP address. Passed to http/https request. Local interface from which the request should be emitted.uri_modifier: Anonymous function taking request (or redirect location if following redirects) URI as an argument and modifying it given logic. It has to return a valid URI string for successful request.decode_response : (or decode) Whether to decode the text responses to UTF-8, if Content-Type header shows a different charset. Defaults to true.parse_response : (or parse) Whether to parse XML or JSON response bodies automagically. Defaults to true. You can also set this to 'xml' or 'json' in which case Needle will only parse the response if the content type matches.output : Dump response output to file. This occurs after parsing and charset decoding is done.parse_cookies : Whether to parse response’s Set-Cookie header. Defaults to true. If parsed, response cookies will be available at resp.cookies.These are basically shortcuts to the headers option described above.
cookies : Builds and sets a Cookie header from a { key: 'value' } object.compressed: If true, sets 'Accept-Encoding' header to 'gzip,deflate', and inflates content if zipped. Defaults to false.username : For HTTP basic auth.password : For HTTP basic auth. Requires username to be passed, but is optional.accept : Sets 'Accept' HTTP header. Defaults to */*.connection: Sets 'Connection' HTTP header. Not set by default, unless running Node < 0.11.4 in which case it defaults to close. More info about this below.user_agent: Sets the 'User-Agent' HTTP header. Defaults to Needle/{version} (Node.js {node_version}).content_type: Sets the 'Content-Type' header. Unset by default, unless you're sending data in which case it's set accordingly to whatever is being sent (application/x-www-form-urlencoded, application/json or multipart/form-data). That is, of course, unless the option is passed, either here or through options.headers. You're the boss.These options are passed directly to https.request if present. Taken from the original documentation:
pfx : Certificate, Private key and CA certificates to use for SSL.key : Private key to use for SSL.passphrase : A string of passphrase for the private key or pfx.cert : Public x509 certificate to use.ca : An authority certificate or array of authority certificates to check the remote host against.ciphers : A string describing the ciphers to use or exclude.rejectUnauthorized : If true, the server certificate is verified against the list of supplied CAs. An 'error' event is emitted if verification fails. Verification happens at the connection level, before the HTTP request is sent.secureProtocol : The SSL method to use, e.g. SSLv3_method to force SSL version 3.family : IP address family to use when resolving host and hostname. Valid values are 4 or 6. When unspecified, both IP v4 and v6 will be used.These options only apply if the follow_max (or follow) option is higher than 0.
follow_set_cookies : Sends the cookies received in the set-cookie header as part of the following request. false by default.follow_set_referer : Sets the 'Referer' header to the requested URI when following a redirect. false by default.follow_keep_method : If enabled, resends the request using the original verb instead of being rewritten to get with no data. false by default.follow_if_same_host : When true, Needle will only follow redirects that point to the same host as the original request. false by default.follow_if_same_protocol : When true, Needle will only follow redirects that point to the same protocol as the original request. false by default.follow_if_same_location : Unless true, Needle will not follow redirects that point to same location (as set in the response header) as the original request URL. false by default.Yes sir, we have it. Needle includes a defaults() method, that lets you override some of the defaults for all future requests. Like this:
needle.defaults({
open_timeout: 60000,
user_agent: 'MyApp/1.2.3',
parse_response: false });
This will override Needle's default user agent and 10-second timeout, and disable response parsing, so you don't need to pass those options in every other request.
Since you can pass a custom HTTPAgent to Needle you can do all sorts of neat stuff. For example, if you want to use the tunnel module for HTTPS proxying, you can do this:
var tunnel = require('tunnel');
var myAgent = tunnel.httpOverHttp({
proxy: { host: 'localhost' }
});
needle.get('foobar.com', { agent: myAgent });
Otherwise, you can use the hpagent package, which keeps the internal sockets alive to be reused.
const { HttpsProxyAgent } = require('hpagent');
needle('get', 'https://localhost:9200', {
agent: new HttpsProxyAgent({
keepAlive: true,
keepAliveMsecs: 1000,
maxSockets: 256,
maxFreeSockets: 256,
scheduling: 'lifo',
proxy: 'https://localhost:8080'
})
});
Unless you're running an old version of Node (< 0.11.4), by default Needle won't set the Connection header on requests, yielding Node's default behaviour of keeping the connection alive with the target server. This speeds up immensely the process of sending several requests to the same host.
On older versions, however, this has the unwanted behaviour of preventing the runtime from exiting, either because of a bug or 'feature' that was changed on 0.11.4. To overcome this Needle does set the 'Connection' header to 'close' on those versions, however this also means that making new requests to the same host doesn't benefit from Keep-Alive.
So if you're stuck on 0.10 or even lower and want full speed, you can simply set the Connection header to 'Keep-Alive' by using { connection: 'Keep-Alive' }. Please note, though, that an event loop handler will prevent the runtime from exiting so you'll need to manually call process.exit() or the universe will collapse.
needle.get('https://api.server.com', { username: 'you', password: 'secret' },
function(err, resp) {
// used HTTP auth
});
Or use RFC-1738 basic auth URL syntax:
needle.get('https://username:password@api.server.com', function(err, resp) {
// used HTTP auth from URL
});
needle.get('other.server.com', { username: 'you', password: 'secret', auth: 'digest' },
function(err, resp, body) {
// needle prepends 'http://' to your URL, if missing
});
var options = {
compressed : true,
follow : 10,
accept : 'application/vnd.github.full+json'
}
needle.get('api.github.com/users/tomas', options, function(err, resp, body) {
// body will contain a JSON.parse(d) object
// if parsing fails, you'll simply get the original body
});
needle.get('https://news.ycombinator.com/rss', function(err, resp, body) {
// you'll get a nice object containing the nodes in the RSS
});
needle.get('http://upload.server.com/tux.png', { output: '/tmp/tux.png' }, function(err, resp, body) {
// you can dump any response to a file, not only binaries.
});
needle.get('http://search.npmjs.org', { proxy: 'http://localhost:1234' }, function(err, resp, body) {
// request passed through proxy
});
var stream = needle.get('http://www.as35662.net/100.log');
stream.on('readable', function() {
var chunk;
while (chunk = this.read()) {
console.log('got data: ', chunk);
}
});
var stream = needle.get('http://jsonplaceholder.typicode.com/db', { parse: true });
stream.on('readable', function() {
var node;
// our stream will only emit a single JSON root node.
while (node = this.read()) {
console.log('got data: ', node);
}
});
// The 'data' element of this stream will be the string representation
// of the titles of all posts.
needle.get('http://jsonplaceholder.typicode.com/db', { parse: true })
.pipe(new JSONStream.parse('posts.*.title'));
.on('data', function (obj) {
console.log('got post title: %s', obj);
});
var data = {
foo: 'bar',
image: { file: '/home/tomas/linux.png', content_type: 'image/png' }
}
needle.post('http://my.other.app.com', data, { multipart: true }, function(err, resp, body) {
// needle will read the file and include it in the form-data as binary
});
needle.put('https://api.app.com/v2', fs.createReadStream('myfile.txt'), function(err, resp, body) {
// stream content is uploaded verbatim
});
var buffer = fs.readFileSync('/path/to/package.zip');
var data = {
zip_file: {
buffer : buffer,
filename : 'mypackage.zip',
content_type : 'application/octet-stream'
}
}
needle.post('http://somewhere.com/over/the/rainbow', data, { multipart: true }, function(err, resp, body) {
// if you see, when using buffers we need to pass the filename for the multipart body.
// you can also pass a filename when using the file path method, in case you want to override
// the default filename to be received on the other end.
});
var data = {
token: 'verysecret',
payload: {
value: JSON.stringify({ title: 'test', version: 1 }),
content_type: 'application/json'
}
}
needle.post('http://test.com/', data, { timeout: 5000, multipart: true }, function(err, resp, body) {
// in this case, if the request takes more than 5 seconds
// the callback will return a [Socket closed] error
});
For even more examples, check out the examples directory in the repo.
To run tests, you need to generate a self-signed SSL certificate in the test directory. After cloning the repository, run the following commands:
$ mkdir -p test/keys
$ openssl genrsa -out test/keys/ssl.key 2048
$ openssl req -new -key test/keys/ssl.key -x509 -days 999 -out test/keys/ssl.cert
Then you should be able to run npm test once you have the dependencies in place.
Note: Tests currently only work on linux-based environments that have
/proc/self/fd. They do not work on MacOS environments. You can use Docker to run tests by creating a container and mounting the needle project directory on/appdocker create --name Needle -v /app -w /app -v /app/node_modules -i node:argon
Written by Tomás Pollak, with the help of contributors.
(c) Fork Ltd. Licensed under the MIT license.
FAQs
The leanest and most handsome HTTP client in the Nodelands.
We found that needle demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
Research
A malicious npm package disguised as a WhatsApp client is exploiting authentication flows with a remote kill switch to exfiltrate data and destroy files.
Security News
PyPI now supports digital attestations, enhancing security and trust by allowing package maintainers to verify the authenticity of Python packages.
Security News
GitHub removed 27 malicious pull requests attempting to inject harmful code across multiple open source repositories, in another round of low-effort attacks.