How to write a connector


New! Want an interactive tutorial on how to create a connector? Try this one:

A connector is a small script that allows to import data from an external website. Each connector is an independent application, managed by the Cozy Collect application.

To protect your data, each connector runs inside a sandbox where all their interactions with your data are under control.

From a technical point of view, connectors are Node.js applications executed inside a container. They communicate with the server using its API, like client side apps, and get an auth token every time they start. They need to register with a manifest, and ask permissions to the user.

To ease the development of a connector, an npm package, named cozy-connector-libs provides a lot of shared libraries. But you may need some other npm packages to help you running your connector:

When the application is started, it also gets some data through environment variables:

But the base connector (require('cozy-konnector-libs').BaseKonnector) in cozy-konnector-libs handles these for you.

The application can access a temporary file system, deleted at the end of its execution. Its logs (standard and error output) are kept by the server.

From the server point of view, each connector is a job run through a trigger.

Let’s create our first connector

The easiest way to create a new connector is to use our template:

git clone
cd cozy-konnector-template
yarn # or npm install

note: the Cozy Team uses yarn, but if you prefer npm, just keep using it, everything should just work.

Then, write your code into index.js and build the application, running yarn build or npm run build.

Collector structure

Basically, a connector is just a function passed to the BaseKonnector constructor, and which returns a promise:

To create the connector, juste create a new instance of BaseKonnector with the function as argument

const {BaseKonnector, saveFiles, request} = require('cozy-konnector-libs')

const rq = request()

module.exports = new BaseKonnector(fields => {
  return rq({
    uri: '',
    qs: {
      q: 'chatons',
      count: 10
  .then(body => {
    let result = []
    if (body && && && {
      result = => ({fileurl:}))
    return result
  .then(entries => saveFiles(entries, fields))

Fetch operations

Every time the connector is run, it will call the function and wait for the resolution of the return promise. This function can then log into the remote site, fetch data and save it.

A basic connector workflow involves: - getting data and storing them into a variable. You can get the data by calling an API, scraping the remote website… - filtering data to remove the ones already present inside the database using filterData - save the filtered data into the database (addData) - save the related files using (saveFiles)

We’ll have a deeper look at this methods below.

Error handling

If your connector hit some issue fetching or saving the data, it can return an error code by throwing it as an error. Some error code are defined inside the Cozy Collect application and will display an explicit error to the user:

Konnector lib

The Cozy Konnector Lib provide some useful methods for common tasks:

Common methods


If you want to access cozy-client-js directly, this method gives you directly an instance of it, initialized according to COZY_URL and COZY_CREDENTIALS environment variable. You can refer to the cozy-client-js documentation for more information.

const {clientClient} = require('cozy-konnector-libs')'my.doctype', ['_id'])


This function filters the passed array from data already present in the cozy Parameters:

const data = [
    name: 'toto',
    height: 1.8
    name: 'titi',
    height: 1.7

return filterData(data, 'io.cozy.height', {
  keys: ['name']
}).then(filteredData => addData(filteredData, 'io.cozy.height'))


This function save the data into the cozy



The goal of this function is to save the given files in the given folder via the Cozy API.


The goal of this function is create or update the given entries according to if they already exist in the cozy or not


The manifest

Each connector is described by a Manifest. This is a JSON file named manifest.konnector at the root of your code folder. It should include the following information:


TODO See documentation of the manifest of an application

The connector parameters are stored in io.cozy.accounts documents, so each connector should get access to this doctype.


How do I scrap a website

Use the request function from cozy-konnector-libs with the proper options

Here’s a sample code that will fetch the login page to get the value of the anti-CSRF token, submit the login form, browse to the bills page and fetch a bill:

const {BaseKonnector, request} = require('cozy-konnector-libs')
const rq = request({
  jar: true, // handle the cookies like a browser
  json: false, // do not try to parse the result as a json document
  cheerio: true // automatically parse the result with [cheerio](
const moment = require('moment')

module.exports = new BaseKonnector(function fetch (fields) {
  return rq("https://login.remote.web")
  .then($ => { // the result is automaticall wrapped with cheerio and you can use it like jQuery
    const form = {
      form: {
        login: fields.login,
        password: fields.password,
        csrf_token: $('[name="csrf_token"]').val(),
    return rq({
      method: 'POST',
  .then($ => rq("https://admin.remote.web/bills"))
  .then($ => {
    return [{date: moment($("#bill_date")), value: $("#bill_value")}]
  .then(entries => addData(entries, 'io.cozy.bills'))


Running in standalone mode

To ease the development, you don’t need a running Cozy server to test your code. We provide a standalone mode, that mocks the server. This mode uses a configuration file to define the environment variables that the server will send to your application in production. So, start by running yarn standalone, a konnector-dev-config.json file will be created and set the parameters your application requires. Then start it with yarn standalone or npm run standalone.

In standalone mode, saving a file will put it at the root of your repository. If you need to query the database, put your mock data into fixture.json. Also, fetched data will just be outputted to the console instead of being sent to the database.