:octocat: Github API Client using Promises or callbacks. Intended for the browser or NodeJS.
JavaScript CoffeeScript Shell HTML
Latest commit ecf6ddb May 26, 2017 @philschatz 🎨 refer to fetch-vcr



gh-board NPM version Downloads build status dependency status dev dependency status code coverage

Try it out in your browser! (REPL)

Octokat.js provides a minimal higher-level wrapper around GitHub's API. It is being developed in the context of an EPUB3 Textbook editor for GitHub and a simple serverless kanban board (demo). It is also used in the official github desktop client!

This package can be used in nodejs or in the browser as an AMD module or using browserify.

Table of Contents


Octokat runs in node or a browser and is available in npm.

npm install --save octokat

Key Features

  • Works in nodejs, an AMD module in the browser, and as a bower library
  • Handles text and binary files
  • Exposes everything available via the GitHub API (repos, teams, events, hooks, emojis, etc.)
  • Supports ETag caching
  • Paged results
  • Node-style callbacks as well as optional Promises (to avoid those debates)
  • 100% of the GitHub API
    • Starring and Following repositories, users, and organizations
    • Editing Team and Organization Membership
    • User/Org/Repo events and notifications
    • Listeners for rate limit changes
    • Public Keys
    • Hooks (commit, comment, etc.)
    • Uses native Promises if available
    • Markdown generation
    • Preview APIs (Deployments, Teams, Licenses, etc)
    • Enterprise APIs

For the full list of supported methods see ./src/grammar/, ./examples/, Travis tests, or the ./test directory.


This library closely mirrors the https://developer.github.com/v3 documentation.

For example:

// `GET /repos/:owner/:repo` in the docs becomes:
octo.repos(owner, repo).fetch()

// `POST /repos/:owner/:repo/issues/:number/comments` becomes:
octo.repos(owner, repo).issues(number).comments.create(params)

The last method should be a verb method. The verb method makes the async call and should either have a callback as the last argument or it returns a Promise (see Promises or Callbacks).

The basic structure of the verb method is:

  • .foos.fetch({optionalStuff:...}) yields a list of items (possibly paginated)
  • .foos(id).fetch(...) yields a single item (issue, repo, user)
  • .foos.create(...) creates a new foo
  • .foos(id).update(...) updates an existing foo
  • .foos(id).add() adds an existing User/Repo (id) to the list
  • .foos(id).remove() removes a member from a list or deletes the object and yields a boolean indicating success
  • .foos.contains(id) tests membership in a list (yields true/false)
  • .foos(id).read() is similar to .fetch() but yields the text contents without the wrapper JSON
  • .foos(id).readBinary() is similar to .read() but yields binary data


Below are some examples for using the library. For a semi-autogenerated list of more examples see ./examples/.


You construct the URL by chaining properties and methods together and an async call is made once a verb method is called (see below).

octo = new Octokat()
repo = octo.repos('philschatz', 'octokat.js')
// Check if the current user is a collaborator on a repo
.then((isCollaborator) => {
  // If not, then star the Repo
  if (!isCollaborator) {
    .then(() => {
      // Done!

Or, update a specific comment:

octo = new Octokat({token: ...})
octo.repos('philschatz', 'octokat.js').issues(1).comments(123123).update({body: 'Hello'})
.then(() => {
  // Done!

Promises or Callbacks

This library supports Node.js-style callbacks as well as Promises.

To use a callback, just specify it as the last argument to a method. To use a Promise, do not specify a callback and the return value will be a Promise.

Example (get information on a repo):

// Using callbacks
octo.repos('philschatz', 'octokat.js').fetch((err, repo) => {
  if (err) console.error(err)
  // Do fancy stuff...

// Using Promises
octo.repos('philschatz', 'octokat.js').fetch()
.then((repo) => {
  // Do fancy stuff
}).then(null, (err) => console.error(err))

Read/Write/Remove a File

To read the contents of a file:

var octo = new Octokat()
var repo = octo.repos('philschatz', 'octokat.js')
repo.contents('README.md').read() // Use `.read` to get the raw file.
.then((contents) => {        // `.fetch` is used for getting JSON

To read the contents of a binary file:

var octo = new Octokat()
var repo = octo.repos('philschatz', 'octokat.js')
repo.contents('README.md').readBinary() // Decodes the Base64-encoded content
.then((contents) => {

To read the contents of a file and JSON metadata:

var octo = new Octokat()
var repo = octo.repos('philschatz', 'octokat.js')
.then((info) => {
  console.log(info.sha, info.content)

To update a file you need the blob SHA of the previous commit:

var octo = new Octokat({token: 'API_TOKEN'})
var repo = octo.repos('philschatz', 'octokat.js')
var config = {
  message: 'Updating file',
  content: base64encode('New file contents'),
  sha: '123456789abcdef', // the blob SHA
  // branch: 'gh-pages'

.then((info) => {
  console.log('File Updated. new sha is ', info.commit.sha)

Creating a new file is the same as updating a file but the sha field in the config is omitted.

To remove a file:

var octo = new Octokat({token: 'API_TOKEN'})
var repo = octo.repos('philschatz', 'octokat.js')
var config = {
  message: 'Removing file',
  sha: '123456789abcdef',
  // branch: 'gh-pages'

.then(() => {
  console.log('File Updated')


All asynchronous methods accept a Node.js-style callback and return a Common-JS Promise.

In a Browser

Create an Octokat instance.

var octo = new Octokat({
  username: "USER_NAME",
  password: "PASSWORD"

var cb = function (err, val) { console.log(val) }

octo.repos('philschatz', 'octokat.js').fetch(cb) // Fetch repo info
octo.me.starred('philschatz', 'octokat.js').add(cb) // Star a repo

Or if you prefer OAuth:

var octo = new Octokat({
  token: "OAUTH_TOKEN"

In a browser using RequireJS

define(['octokat'], (Octokat) => {
  var octo = new Octokat({
    username: "YOU_USER",
    password: "YOUR_PASSWORD"

In Node.js

Install instructions:

npm install octokat --save
var Octokat = require('octokat')
var octo = new Octokat({
  username: "YOU_USER",
  password: "YOUR_PASSWORD"

// You can omit `cb` and use Promises instead
var cb = function (err, val) { console.log(val) }

octo.repos('philschatz', 'octokat.js').fetch(cb)    // Fetch repo info
octo.me.starred('philschatz', 'octokat.js').add(cb) // Star a repo
octo.me.starred('philschatz', 'octokat.js').remove(cb) // Un-Star a repo

Using bower

This file can be included using the bower package manager:

bower install octokat --save


This is all you need to get up and running:

<script src="../dist/octokat.js"></script>
  var octo = new Octokat()
  octo.zen.read((err, message) => {
    if (err) { throw new Error(err) }

Advanced Uses


GitHub provides URL patterns in its JSON responses. These are automatically converted into methods. You can disable this by setting disableHypermedia: true in the options when creating a new Octokat(...).

For example:

octo.repos('philschatz', 'octokat.js').fetch()
.then((repo) => {
  // GitHub returns a JSON which contains something like compare_url: 'https://..../compare/{head}...{base}
  // This is converted to a method that accepts 2 arguments
  repo.compare(sha1, sha2).fetch()
  .then((comparison) => console.log('Done!'))

Paged Results

If a .fetch() returns paged results then nextPage(), previousPage(), firstPage() and lastPage() are added to the returned Object. For example:

octo.repos('philschatz', 'octokat.js').commits.fetch()
.then((someCommits) => {
  .then((moreCommits) => {
    console.log('2nd page of results', moreCommits)

As standard with the Github API, passing a per_page parameter allows you to control the number of results per page. For example:

octo.repos('philschatz', 'octokat.js').issues.fetch({per_page: 100})

Preview new APIs

Octokat will send the Preview Accept header by default for several Preview APIs.

If you want to change this behavior you can force an acceptHeader when instantiating Octokat.

For example:

var octo = new Octokat({
  token: 'API_TOKEN',
  acceptHeader: 'application/vnd.github.cannonball-preview+json'

Enterprise APIs

To use the Enterprise APIs add the root URL when instantiating Octokat:

var octo = new Octokat({
  token: 'API_TOKEN',
  rootURL: 'https://example.com/api/v3'

Using EcmaScript 6 Generators

This requires Node.js 0.11 with the --harmony-generators flag:

var Octokat = require('octokat')
var octo = new Octokat()

var zen  = yield octo.zen.read()
var info = yield octo.repos('philschatz', 'octokat.js').fetch()


Uploading Releases

Uploading release assets requires a slightly different syntax because it involves setting a custom contentType and providing a possibly binary payload.

To upload (tested using nodejs) you can do the following:

var contents = fs.readFileSync('./build.js')

.then((release) => {

  release.upload('build.js', 'application/javascript', contents)
    .then((resp) => {
      // Success!

Parsing JSON

If you are using webhooks, the JSON returned by GitHub can be parsed using octo.parse(json) to yield a rich object with all the methods Octokat provides.

octo.parse(json) is asynchronous and can take either a callback or returns a promise.

Using URLs Directly

Instead of using Octokat to construct URLs, you can construct them yourself and still use Octokat for sending authentication information, caching, pagination, and parsing Hypermedia.

// Specify the entire URL

// Or, just the path
octo.fromUrl('/repos/philschatz/octokat.js/issues').fetch({state: 'open'}, cb)

If the URL is a Hypermedia Template then you can fill in the fields by passing them in as an additional argument.

params = {
  owner: 'philschatz'
  repo: 'octokat.js'
  name: 'dist.js'
octo.fromUrl('https://uploads.github.com/repos/{owner}/{repo}/releases{/id}/assets{?name}', params)
// returns https://uploads.github.com/repos/philschatz/octokat.js/releases/assets?name=dist.js


  • Run npm install
  • Run npm test to run Mocha tests for Node.js and the browser
  • Run grunt dist to generate the files in the ./dist directory

The unit tests are named to illustrate examples of using the API. See Travis tests or run npm test to see them.

fetch-vcr is used to generate recorded HTTP fixtures from GitHub. If you are adding tests be sure to include the updated fixtures in the Pull Request.


  • Add Option for Two factor authentication
  • Add option to pass header as cache control: no cache