The client-side agent is a high-performance JavaScript agent that collects multiple device and browser signals and sends them to the FingerprintJS Pro API for processing and identification.

Installing the agent

There are various ways to install the agent. No matter what you choose, the usage will be the same.

Global variable

This is the easiest way to start. Add the following HTML code to your pages:

<script>
  // Initialize an agent at application startup.
  const fpPromise = FingerprintJS.load({ token: 'your-browser-token' });

  // Get the visitor identifier when you need it.
  fpPromise
    .then(fp => fp.get())
    .then(result => console.log(result.visitorId));
</script>
<script
  async
  src="//cdn.jsdelivr.net/npm/@fingerprintjs/[email protected]/dist/fp.min.js"
  onload="initFingerprintJS()"
></script>

You can get a token at dashboard.fingerprintjs.com.

ECMAScript module

We recommend this installation method. First install the agent NPM package:

npm install @fingerprintjs/fingerprintjs-pro

yarn add @fingerprintjs/fingerprintjs-pro

Then import the package into your code:

import FingerprintJS from '@fingerprintjs/fingerprintjs-pro'

// Initialize an agent at application startup.
const fpPromise = FingerprintJS.load({ token: 'your-browser-token' })

// Get the visitor identifier when you need it.
fpPromise
  .then(fp => fp.get())
  .then(result => console.log(result.visitorId))

EcmaScript module comes with a TypeScript declaration for superior development experience. The module can be used with various module bundlers such as Webpack, Rollup.js or Browserify.

UMD

It's designed for module loaders like RequireJS. Use the following JS code:

require(
  ['//cdn.jsdelivr.net/npm/@fingerprintjs/[email protected]/dist/fp.umd.min.js'],
  (FingerprintJS) => {
    // Initialize an agent.
    const fpPromise = FingerprintJS.load({ token: 'your-browser-token' })

    // Get the visitor identifier when you need it.
    fpPromise
      .then(fp => fp.get())
      .then(result => console.log(result.visitorId))
  }
)

CommonJS

It's very similar to ECMAScript module, but it's legacy. First install the agent NPM package:

npm install @fingerprintjs/fingerprintjs-pro

yarn add @fingerprintjs/fingerprintjs-pro

Then import the package in your code:

const FingerprintJS = require('@fingerprintjs/fingerprintjs-pro')

// Initialize an agent at application startup.
const fpPromise = FingerprintJS.load({ token: 'your-browser-token' })

// Get the visitor identifier when you need it.
fpPromise
  .then(fp => fp.get())
  .then(result => console.log(result.visitorId))

Agent initialization

You should create an agent when your page or application starts:

const fpPromise = FingerprintJS.load({ /* ... */ })

// fpPromise will resolve with `fp` object that you can use to call `fp.get()`

To configure the agent, pass parameters into the “load” function. The table below summarizes the parameters available. Below the table are detailed descriptions of each parameter

Full list of parameters:

token default: no, required: yes

Browser token to authenticate the agent with the server API. You can get one at dashboard.fingerprintjs.com.

Example usage:

const fpPromise = FingerprintJS.load({ token: "dXk12f271339d9ag93d" })

regiondefault: 'us', required: false

Available values: us and eu

The region of the FingerprintJS Pro servers you want to send data to. The default is us, which sends data to the N. Virginia server location. This parameter is ignored when endpoint and tlsEndpoint parameters are used.

Example:

const fpPromise = FingerprintJS.load({ region: "us" })

endpointdefault: no, required: false

This parameter should only be used with the Subdomain integration. Specify your custom endpoint here.

tlsEndpointdefault: no, required: false, since: v3.1.0

Your custom TLS endpoint URL address.

storageKeydefault: '_vid', required: false

Name of key to store data in visitor browsers. The data is stored in cookies and local storage. You shouldn't change this parameter once your code runs in production because a change will cause the data in visitor browsers to be purged which will decrease identification accuracy.

Visitor identification

Once an agent instance is initialized, you can start getting visitor identification. You are free to identify at any moment of time. You can configure FingerprintJS Pro to receive the identification data in the browser, in the server (via a webhook), or both.

Go to the extendedResult section to see the browser result format. The visitorId field will contain an empty string when the visitor can't be identified, for example when it's a crawler bot.

Below are a few examples of different approaches for identifying.

Automatic mode

In many cases you will use the automatic mode, because it's the simplest way to get started. The agent will get the data on every page load. You don't need to do anything else. Just add the call fp.get() as soon as the agent is loaded:

const fpPromise = FingerprintJS.load({ /* ... */ })
const fp = await fpPromise
const result = await fp.get()
console.log(result.visitorId)

You can use both async/await and promises to get a visitor identifier. Example of using promises:

const fpPromise = FingerprintJS.load({ /* ... */ })
fpPromise
    .then(fp => fp.get())
    .then(result => console.log(result.visitorId))

When get() is called, a FingerprintJS server will send a webhook request to your server.

Manual mode

Manual mode provides more granular control over how the identification requests are processed. When you actually need to get visitor data (on an Ad click for example), call the .get() method to get the data.

const fpPromise = FingerprintJS.load({ /* ... */ })

document.querySelector('#ad').addEventListener('click', async () => {
    const fp = await fpPromise
    const result = await fp.get()
    console.log(result.visitorId)
})

const fpPromise = FingerprintJS.load({ /* ... */ })

document.querySelector('#ad').addEventListener('click', () => {
  fpPromise
    .then(fp => fp.get())
    .then(result => console.log(result.visitorId))
})

Search bots

The PRO API recognizes most well-known search bots (Google, Bing, Yahoo, Baidu, DuckDuckGo, and others) and treats them differently. Whenever a page is loaded by a search bot, an identification event is processed server-side, like it would with a normal visitor. The visitorId field will contain an empty string, and the bot.safe value will be true (available when an extended result is requested). Crawl bot detection example:

const result = await fp.get({ extendedResult: true })
if (!result.visitorId && result.bot && result.bot.safe) {
  console.log('This is a crawl bot')
}

Parameters reference

tag

tag is a customer-provided value or an object that will be saved together with the visit and will be returned back to you in a webhook message or when you search for the visit in the visitor API. You may want to use the tag value to be able to associate a visit event with information from your current visitor's web request context or add any other information that is convenient for you.

Is the tag field required?
No, it is optional.

What values can be used as a tag?
Any value that you need when processing a webhook event or that is useful when searching for visits in our visitor API. Scalar, string or object values can be sent.
For example you can pass a yourCustomId as a tag and then get this yourCustomId back in the webhook, associated with a visitorId.

Examples of using the tag option:

// simple, scalar values
const yourCustomId = 2718281828
fp.get({ tag: yourCustomId })
// or a string value
const requestType = "signup"

// Example of sending and receiving the tag value
fp.get({ tag: yourCustomId })
// Webhook request
{
  "visitorId": "nTxJ28Gt4CegeAwdWELL",
  "tag": { "tag": 2718281828 },
  // ...
}
// Note that tags are always objects, even when you use
// a regular scalar value. 
// If you provide a scalar value, it will converted to an object
// with `tag` attribute.

// You can also use a JavaScript object as a tag value 
// and it will be returned back to you in a webhook.
// 
// NOTE: nested objects are not supported in tag objects.
const requestMetadata = {
  yourCustomId: 2718281828, 
  requestType: "signup"
}
// Making an API call and receiving a response via a webhook request
fp.get({ tag: requestMetadata })
// Webhook request
{
  "visitorId": "nTxJ28Gt4CegeAwdWELL", 
  "tag": { "yourCustomId": 2718281828, "requestType": "signup" },
  // ...
}

linkedId

linkedId is a way of linking current identification event with a custom identifier. This can be helpful to be able to filter API visit information later.

Some examples of how you can use the linkedId:

1. Link your website userId with a FPJS visitorId to be able to analyze this data.
2. Link your e-commerce orderId with a FPJS visitorId to be able to see the orders a visitor made.
3. Link a survey ID with a visitorId to ensure that the same visitor does not participate in a single survey multiple times.
4. Limit one visitor per item or one visitor per discount coupon.

// Making an API call
// Sending a linkedId value is helpful to filter by this value via server API
var orderId = 3936532456
fp.get({ linkedId: orderId })
// linkedId will be saved with this event and will be available to be queried later.

A linkedId looks similar to tag. When should you use each?
A tag is an object that is "passed-through" and sent back to your webhook endpoint. A tag can contain an arbitrary number of attributes of any type (except for nested objects). You can use tags to store them in your database and use for any purpose. Tag objects are meant to be consumed and processed by you eventually. You can think of a tag object as metadata.
A linkedId is a scalar, string identifier that gets indexed by FPJS platform and can be used to filter visits efficiently. It's a way to create a custom index on your data and use that index to efficiently filter visits when querying them through the API. Objects or non-string values are not supported.

ipResolution

ipResolution is an option that controls the geolocation accuracy.

extendedResult

Two types of responses are supported: "default" and "extended". You don't need to pass any parameters to get the "default" response. "Extended" is an extended result format and can be requested using the extendedResult: true parameter.

Response object format is provided below. Every request has a unique requestId that is returned in the response and is also available in API to query later. It is a convenient way to get specific identification events in the API.

fp.get()...
// response:
{
  // requestId is different for every request
  // it's a request identifier
  "requestId": "8nbmT18x79m54PQ0GvPq",
  // visitorId is the main browser identifier
  "visitorId": "2JGu1Z4d2J4IqiyzO3i4",
  // if true, this visitor was found and visited before
  // if false, this visitor wasn't found and probably didn't visit before
  "visitorFound": true
}

fp.get({ extendedResult: true })...
// response:
{
  "requestId": "8nbmT18x79m54PQ0GvPq",
  "visitorId": "2JGu1Z4d2J4IqiyzO3i4",
  "visitorFound": true,
  "ip": "185.230.125.20",
  "ipLocation": {
    "accuracyRadius": 10,
    "latitude": 47.3925,
    "longitude": 8.4546,
    "postalCode": "8010",
    "timezone": "Europe/Zurich",
    "city": {
      "name": "Zurich"
    },
    "continent": {
      "code": "EU",
      "name": "Europe"
    },
    "country": {
      "code": "CH",
      "name": "Switzerland"
    },
    "subdivisions": [
      {
        "isoCode": "ZH",
        "name": "Zurich"
      }
    ]
  },
  "browserName": "Chrome",
  "browserVersion": "75.0.3770",
  "os": "Mac OS X",
  "osVersion": "10.14.5",
  "device": "Other",
  "bot": { probability: 1, safe: true },
  "incognito": false
}

timeout

Controls client-side timeout. Client timeout controls total time (both client-side and server-side) that any identification event is allowed to run. It doesn't include time when the page is in background (not visible) because browser may suspend identification process in background. By default it's 10 seconds. You can control the client-side timeout using timeout configuration key. The timeout option is in milliseconds. Example usage:

// A timeout of 20 seconds
// An example of the client-side timeout handling
try {
  const result = await fp.get({ timeout: 20000 })
  // ...
} catch (error) {
  if (error.message === FingerprintJS.ERROR_CLIENT_TIMEOUT) {
    console.log("A timeout of 20 seconds exceeded")
  }
}

disableTls

Set to true to disable the extra TLS request.

Error handling

JavaScript agent get method returns a promise which will be rejected in case of an error. Below table summarizes possible types of errors that can lead to result promise rejections. See detailed information about the errors below the table.

With the exception of ERROR_CLIENT_TIMEOUT, ERROR_NETWORK_CONNECTION and ERROR_NETWORK_ABORT, all the errors described above will include the requestId field. The get method can also throw other unexpected errors, they should be treated as agent bugs.

Error handling example:

const fpPromise = FingerprintJS.load({ /* ... */ })

try {
  const fp = await fpPromise
  const result = await fp.get()
} catch (error) {
  switch (error.message) {
    case FingerprintJS.ERROR_GENERAL_SERVER_FAILURE:
      console.log('Unknown server error. Request id:', error.requestId)
      break
    case FingerprintJS.ERROR_CLIENT_TIMEOUT:
      console.log('Identification time limit of 10 seconds is exceeded')
      break
    default:
      console.log('Other error')
  }
}

The connection error occurs when the agent fails to connect to the server API. This can happen during a network outage or if a browser plugin blocks networking requests (e.g. AdBlock).

Timeouts

Two types of timeouts are possible: a server timeout and a client timeout.

Server timeout

The server timeout is fixed at 10 seconds of server-side processing time. If server-side processing exceeds 10 seconds for any reason, the promise will be rejected.

Client timeout

Client timeout controls total time (both client-side and server-side) that any identification event is allowed to run. By default it's 10 seconds. Note that even if the client-side timeout is exceeded, the server-side request can still be running, its results will just be discarded.

Rate limiting

Every account has a default rate limit of 10 requests / second. If you need to increase that value, email as at [email protected]. Whenever that rate limit is exceeded, a request is throttled and a special error is emitted.

Supported browsers

The JS agent supports all popular browsers. We use the following terminal command to decide which browsers to support:

npx browserslist "cover 95% in us, not IE < 10"

See more details and learn how to run the agent in old browsers in the browser support guide.