🚀 Supporting request forwarding and Promise based HTTP client for all JavaScript runtimes.
JavaScript HTML TypeScript
Switch branches/tags
Nothing to show
Clone or download
Latest commit e375c3c Jul 14, 2018
Failed to load latest commit information.
demon fix #67 May 21, 2018
dist fix #65, #92, #90, #83, #66 Jul 14, 2018
imgs update doc Mar 23, 2018
src fix #65, #92, #90, #83, #66 Jul 14, 2018
test fix #65, #92, #90, #83, #66 Jul 14, 2018
.gitignore build Jun 28, 2018
.npmignore fix #67 May 21, 2018
.travis.yml add group test May 2, 2018
CHANGELOG.md add lock/unlock API for interceptors Apr 20, 2018
LICENSE add license Apr 20, 2018
README-CH.md fix #65, #92, #90, #83, #66 Jul 14, 2018
README.md fix #65, #92, #90, #83, #66 Jul 14, 2018
build.js add group test May 2, 2018
fly.png modify doc Sep 24, 2017
index.d.ts fix #65, #92, #90, #83, #66 Jul 14, 2018
index.js Support execute async task in interceptors Apr 19, 2018
karma.conf.js add group test May 2, 2018
node-index.js 1. Optimize project structure Oct 26, 2017
package.json fix #65, #92, #90, #83, #66 Jul 14, 2018
upload.txt fix #56 Apr 26, 2018
v.png add browser entry Sep 21, 2017



npm version build status coverage size platform


Supporting request forwarding and Promise based HTTP client for all JavaScript runtimes.

Chinese documentation : 中文文档

Browser Support

Chrome Firefox Safari Opera Edge IE
> 8

Other Platforms Support

node logo Mini Program logo mpvue logo weex logo quick app logo

Currently the platforms that fly.js supported include Node.jsWeChat Mini ProgramWeexReact NativeQuick App and the browers, all JavaScript runtimes of these are different. More platforms are supporting...


  1. Supports the Promise API
  2. Make XMLHttpRequests from the browser,light-weight and very light-weight
  3. Supports various JavaScript runtimes
  4. Supports request and response interceptors。
  5. Automatic transforms for JSON data。
  6. Supports switching the underlying Http Engine, easy to adapt to various JavaScript Runtime.
  7. Supports global Ajax interception on browser
  8. Supports request forwarding in hybrid applications

Positioning & target

Fly.is locates to be the ultimate solution for Javascript http requests. That is to say, in any environment that can execute Javascript, as long as it has the ability to access the network, Fly can run on it and provide unified APIs. At the same time, keep lightweight on the browser side.


You can find the Fly documentation on the offical website.



Using npm

npm install flyio

Using CDN (on browsers)

<script src="https://unpkg.com/flyio/dist/fly.min.js"></script>

UMD (on browsers)


Require flyio

The entry files of different JavaScript Runtimes may be different, you can refer on the below, but the entries of Browsers, Node, and React Native are the same。each requirement ways of these platforms are as follows:

  1. Requiring On browsers、Node、React Native

    //for Browsers, Node, and React Native
    var fly=require("flyio")

    The above method requires the default instance of Fly, and you can also create an Fly instance by yourself:

    // for browsers and React Native
    var Fly=require("flyio/dist/npm/fly")
    // for Node
    //var Fly=require("flyio/src/node")
    var fly=new Fly;
  2. Requiring on WeChat Mini Program

    var Fly=require("flyio/dist/npm/wx") 
    var fly=new Fly

    If your project does not use NPM to manage dependencies, you can download the source code directly to your project 。The download links are wx.js or wx.umd.min.js . Download any one, save it to your project directory (assuming named "lib") , and then requiring:

    var Fly=require("../lib/wx") 
    var fly=new Fly; 
  3. Requiring on Quick App

    On Quick App, Fly relies on the fetch module, which needs to be defined in the manifest.json first:

      "features": [
        {"name": "system.fetch"}

    And then create Fly instance:

     var fetch = require("@system.fetch")
     var Fly=require("flyio/dist/npm/hap")
     var fly=new Fly(fetch)
  4. Requiring on Weex

    var Fly=require("flyio/dist/npm/weex")
    var fly=new Fly

After the requirement, you can make the global configuration and add the interceptor to fly , and peform network request with fly.


The following example, if not specified, can be executed in all JavaScript Runtimes.

Performing a GET request

var fly=require("flyio")
// Make a request for a user with a given ID,  and the parameter is passed directly in URL
  .then(function (response) {
  .catch(function (error) {

// The parameter is passed by a object
fly.get('/user', {
      id: 133
  .then(function (response) {
  .catch(function (error) {

Performing a POST request

fly.post('/user', {
    name: 'Doris',
    age: 24
  .then(function (response) {
  .catch(function (error) {

Performing multiple concurrent requests

function getUserRecords() {
  return fly.get('/user/133/records');

function getUserProjects() {
  return fly.get('/user/133/projects');

fly.all([getUserRecords(), getUserProjects()])
  .then(fly.spread(function (records, projects) {
    // Both requests are now complete

Performing the request by request

    timeout:5000 // Set timeout to 5 seconds
.then(d=>{ console.log("request result:",d)})
.catch((e) => console.log("error", e))

Sending URLSearchParams

const params = new URLSearchParams();
params.append('a', 1);
.then(d=>{ console.log("request result:",d)})

Note that URLSearchParams is not supported by all browsers (see caniuse.com), but there is a polyfill available (make sure to polyfill the global environment).

Sending FormData

 var formData = new FormData();
 var log=console.log
 formData.append('username', 'Chris');

Note that FormData is not supported by all browsers (see caniuse.com), but there is a polyfill available (make sure to polyfill the global environment). And there are some differences on the way to support formData in node environment. Please click here for more details

Requesting Stream

  //d.data is a ArrayBuffer instance

In browser, the value of responseType can be one of "arraybuffer" or "blob". In node, you can just set it as "stream"


You can intercept requests or responses before they are handled by then or catch.

// Add a request interceptor
    // Do something before request is sent
    // Complete the request with custom data
    // return Promise.resolve("fake data")

// Add a response interceptor
    (response) => {
      // Do something with response data .
      // Just return the data field of response
      return response.data
    (err) => {
      // Do something with response error
      //return Promise.resolve("ssss")

The structures of the request object in request interceptor.

  baseURL,  //base url
  body, // request parameters
  headers, //custom request headers
  method, // http request method
  timeout, // request time
  url, // request url (or relative path)
  withCredentials, // determine whether sending thirdparty cookies in cross-domain request
  ... // custom field defined in options

The structures of the response object in response interceptor.

  data, //response data
  engine, //http engine,In browser,it's a instance of XMLHttpRequest.
  headers, //response headers
  request  //the origin request object

Remove interceptors

If you may need to remove an interceptor later, just set it to null.


Perform an async task in interceptors

Now, you can perform async task in interceptors !

Let's see an example:

because of security reasons, we need all the requests to set up a csrfToken in the header, if csrfToken does not exist, we need to request a csrfToken first, and then perform the network request, because the request csrfToken progress is asynchronous, so we need to execute this async request in request interceptor. the code is as follows:

var csrfToken="";
var tokenFly=new Fly();
var fly=new Fly();
fly.interceptors.request.use(function (request) {
    //if csrfToken does not exist, we need to request a csrfToken first
    if(!csrfToken) {
         // locking the current instance, let the incomming request task enter a 
         // queue before they enter the request interceptors.
         //Using  another fly instance to request csrfToken.
         //If use the same fly instance, there may lead a infinite loop:
         //(The request will go to the interceptor first, and then
         //enter the interceptor again when launching the new request 
         //in the interceptor....)
         return tokenFly.get("/token").then((d)=>{
            //only return the origin `request` object can make the http request continue.
            // otherwise, the return data will be teated as "response" data.
            return request 
           //fly.clear(); //clear the request queue
           // unlock the current instance, flush the request queue.   
    }else {
        request.headers["csrfToken"]= csrfToken;
        //This line can be omitted. 
        //If the interceptor doesn't return value, `request` will be used by default.
        return request 


  1. The current fly instance will be locked when call fly.lock() . Once the fly instance is locked, the incomming request task maked by it will be hang up and enter a queue before they enter the request interceptors, you can call fly.unlock() to continue the requests or call fly.clear() to cancel the requests in the queue.
  2. Only when you return the request object passed by interceptor at the final , the origin http request will be continued.

And you can also make an async task in the response interceptor. More information about interceptors and examples refer to flyio interceptor.

Error handling

If the request fails, catch will be called; the error object's fields as follows :

  message:"Not Find 404", //error description
  status:404, // error code
  request:{...}, //the request info
  response:{...}, //the error response,
  engine:{}, // The Http engine, it's a instance of XmlHttpRequest in browser.
Code Description
0 network error
1 timeout
2 The file was downloaded successfully, but the save failed. This error only occurred in the node environment
>=200 http status code


  .catch(function (error) {
    console.log('Error', error.message);
    console.log('Error engine:', error.engine)
    // The request was made but no response was received
    // `error.request` holds the request info.
    console.log('Error request info:',error.request);
    if (error.response) {
      // The request was made and the server responded with a status code
      // that falls out of the range of 2xx and 304

Request configuration options

Configurable options

 // `method` is the request method to be used when making the request
  method: 'get', // default
  // `headers` are custom headers to be sent
  // `baseURL` will be prepended to `url` unless `url` is absolute.
  // It can be convenient to set `baseURL` for an instance of fly to pass relative URLs
  // to methods of that instance.
  // `timeout` specifies the number of milliseconds before the request times out.
  // If the request takes longer than `timeout`, the request will be aborted.
  timeout:0,  //default
  // `parseJson` indicates whether or not it is  automatically converted response 
  //  data , the Content-Type of which is application/json, to JSON object,  
  // `withCredentials` indicates whether or not cross-site Access-Control requests
  // should be made using credentials
  withCredentials: false, // default

Configuration supports instance level configuration and single request configuration

Instance level

Instance level configuration can be applied to all requests initiated by the current Fly instance:

// Add common headers
// Set timeout
// Set base url

Single request configuration

Single request configuration is valid only once:

    timeout:5000 //超时设置为5s

If the single configuration and the instance configuration conflict, the single request configuration is preferred .


fly.get(url, data, options)

fly.post(url, data, options)

In browser environment, the type of data can be:


options is the request configuration。

fly.request(url, data, options)

When you use this API, you should explicitly specify the method of options, and the default method is get

// Perform a get request
fly.request("/user/8" null, {method:"get"})
// Perform a delete request
fly.request("/user/8/delete", null, {method:"delete"})
// Perform a put request
fly.request("/user/register", {name:"doris"}, {method:"PUT"})

This API is suitable for use in RESTful API scenarios. For convenience, the corresponding alias methods are provided alias

fly.put(url, data, options)




Helper functions for dealing with concurrent requests.


The parameter of all is an array of promise, then will be called only when all requests are successful, andcatch will be called as long as one fails.

Using application/x-www-form-urlencoded format

By default, fly serializes JavaScript objects to JSON. To send data in the application/x-www-form-urlencoded format instead, you can use one of the following options.

By specifying content-type request header

When you specify content-type value as "application/x-www-form-urlencoded", fly will encode the request data automatically, for example:


This method is universal, which means that it can work well in any JavaScript runtime. There are also some other ways for some specific platforms:

Other ways


In a browser, you can use the URLSearchParams API as follows:

var params = new URLSearchParams();
params.append('param1', 'value1');
params.append('param2', 'value2');
fly.post('/foo', params);

Note that URLSearchParams is not supported by all browsers (see caniuse.com), but there is a polyfill available (make sure to polyfill the global environment).

Alternatively, you can encode data using the qs library:

var qs = require('qs');
fly.post('/foo', qs.stringify({ 'bar': 123 }));


In node.js, you can use the querystring module as follows:

var querystring = require('querystring');
fly.post('http://something.com/', querystring.stringify({ foo: 'bar' }));

You can also use the qs library.


Fly depends on a native ES6 Promise implementation to be supported. If your environment doesn't support ES6 Promises, you can polyfill.


fly includes TypeScript definitions.

import fly from "flyio"

Creating an instance

You can create a new instance of Fly , and then apply different configurations:

// Notice require path is  "flyio/dist/npm/fly"
var  Fly=require("flyio/dist/npm/fly")
var nFly=new Fly();
// Apply different configurations

Http Engine

Fly introduces the concept of Http Engine, and Http Engine is the engine that really initiates http requests. This is typically XMLHttpRequest in browser environment, and in Node environments, any module or library that can initiate a network request can be implemented. Fly can switch the Http Engine freely. In fact, Fly is implemented by switching Http Engine, supporting both the browser environment and the Node environment.However, Http Engine is not limited to Node and browser environment, also can be Android, IOS, electron and so on, it is because of these, Fly has a very unique and powerful feature "request forwarding". Based on the request forwarding, we can implement some useful functions, such as redirecting all the HTTP requests of the hybrid application to Native, and then complete the network request on natvie side. The benefit of this is that we can perform unified certificate validation, cookie management, access control and son on on natvie. More details click here http engine .

Global Ajax interception

In browsers, you can intercept global Ajax requests by replacing XMLHttpRequest with Fly engine, regardless of what network library the upper layer uses. More details click here Ajax hook

Using in Node

Whether in browser environment or in Node environment, Fly provides a unified Promise API in the upper layer. This means that, regardless of whether you are in web development or node development, you can perform HTTP requests in the same way. However, because of the difference of node and browser environment, under the environment of Node, Fly in addition to basic API support, also provides some enhancements to the API, the API mainly involves the file download, file upload, HTTP agents and other powerful features, please refer to Node enhanced API .


In browser environment, the size of a library is very important. In this regard, Fly is doing a good job, on the on hand Fly maintains the powerful function, on the other hand, Fly control its body to the minimum. Min is only about 4.6K, and GZIP is less than 2K after compression, and the volume is 1/4 of Axios。

Project structure

Learn the project structure of Fly.js please refer to : Fly.js source structure


Welcome stars 。