webRequest.onBeforeSendHeaders

This event is triggered before sending any HTTP data, but after all HTTP headers are available. This is a good place to listen if you want to modify HTTP request headers.

To have the request headers passed into the listener along with the rest of the request data, pass "requestHeaders" in the extraInfoSpec array.

To modify the headers synchronously: pass "blocking" in extraInfoSpec, then in your event listener, return a BlockingResponse with a property named requestHeaders, whose value is the set of request headers to send.

To modify the headers asynchronously: pass "blocking" in extraInfoSpec, then in your event listener, return a Promise which is resolved with a BlockingResponse.

If you use "blocking", you must have the "webRequestBlocking" API permission in your manifest.json.

It is possible for extensions to conflict here. If two extensions listen to onBeforeSendHeaders for the same request, then the second listener will see modifications made by the first listener, and will be able to undo any changes made by the first listener. For example, if the first listener adds a Cookie header, and the second listener strips all Cookie headers, then the first listener's modifications will be lost. If you want to see the headers that are actually sent, without the risk that another extension will subsequently alter them, use onSendHeaders, although you can't modify headers on this event.

Not all headers actually sent are always included in requestHeaders. In particular, headers related to caching (for example, Cache-Control, If-Modified-Since, If-None-Match) are never sent. Also, behavior here may differ across browsers.

According to the specification, header names are case-insensitive. This means that to match a particular header, the listener should lowercase the name before comparing it:

for (const header of e.requestHeaders) {
  if (header.name.toLowerCase() === desiredHeader) {
    // process header
  }
}

The browser preserves the original case of the header name as generated by the browser. If the extension's listener changes the case, this change will not be kept.

Syntax

browser.webRequest.onBeforeSendHeaders.addListener(
  listener,             //  function
  filter,               //  object
  extraInfoSpec         //  optional array of strings
)
browser.webRequest.onBeforeSendHeaders.removeListener(listener)
browser.webRequest.onBeforeSendHeaders.hasListener(listener)

Events have three functions:

addListener(listener, filter, extraInfoSpec): Adds a listener to this event.
removeListener(listener): Stop listening to this event. The listener argument is the listener to remove.
hasListener(listener): Check whether listener is registered for this event. Returns true if it is listening, false otherwise.

addListener syntax

Parameters

listener

The function called when this event occurs. The function is passed this argument:

details: object. Details of the request. This includes request headers if you have included "requestHeaders" in extraInfoSpec. See the details section for more information.

Returns: webRequest.BlockingResponse. If "blocking" is specified in the extraInfoSpec parameter, the event listener should return a BlockingResponse object, and can set its requestHeaders property.

filter

webRequest.RequestFilter. A set of filters that restricts the events that is sent to this listener.

extraInfoSpec Optional

array of string. Extra options for the event. You can pass any of the following values:

"blocking": make the request synchronous, so you can modify request headers
"requestHeaders": include the request headers in the details object passed to the listener

Additional objects

details

cookieStoreId

string. If the request is from a tab open in a contextual identity, the cookie store ID of the contextual identity. See Work with contextual identities for more information.

documentUrl

string. URL of the document in which the resource will be loaded. For example, if the web page at "https://example.com" contains an image or an iframe, then the documentUrl for the image or iframe will be "https://example.com". For a top-level document, documentUrl is undefined.

frameAncestors

array. Contains information for each document in the frame hierarchy up to the top-level document. The first element in the array contains information about the immediate parent of the document being requested, and the last element contains information about the top-level document. If the load is actually for the top-level document, then this array is empty.

url: string. The URL that the document was loaded from.
frameId: integer. The frameId of the document. details.frameAncestors[0].frameId is the same as details.parentFrameId.

frameId

integer. Zero if the request happens in the main frame; a positive value is the ID of a subframe in which the request happens. If the document of a (sub-)frame is loaded (type is main_frame or sub_frame), frameId indicates the ID of this frame, not the ID of the outer frame. Frame IDs are unique within a tab.

incognito

boolean. Whether the request is from a private browsing window.

method

string. Standard HTTP method: for example, "GET" or "POST".

originUrl

string. URL of the resource which triggered the request. For example, if "https://example.com" contains a link, and the user clicks the link, then the originUrl for the resulting request is "https://example.com".

The originUrl is often but not always the same as the documentUrl. For example, if a page contains an iframe, and the iframe contains a link that loads a new document into the iframe, then the documentUrl for the resulting request will be the iframe's parent document, but the originUrl will be the URL of the document in the iframe that contained the link.

parentFrameId

integer. ID of the frame that contains the frame which sent the request. Set to -1 if no parent frame exists.

proxyInfo

object. This property is present only if the request is being proxied. It contains the following properties:

host

string. The hostname of the proxy server.

port

integer. The port number of the proxy server.

type

string. The type of proxy server. One of:

"http": HTTP proxy (or SSL CONNECT for HTTPS)
"https": HTTP proxying over TLS connection to proxy
"socks": SOCKS v5 proxy
"socks4": SOCKS v4 proxy
"direct": no proxy
"unknown": unknown proxy

username

string. Username for the proxy service.

proxyDNS

boolean. True if the proxy will perform domain name resolution based on the hostname supplied, meaning that the client should not do its own DNS lookup.

failoverTimeout

integer. Failover timeout in seconds. If the proxy connection fails, the proxy will not be used again for this period.

requestHeaders Optional

webRequest.HttpHeaders. The HTTP request headers that will be sent with this request.

requestId

string. The ID of the request. Request IDs are unique within a browser session, so you can use them to relate different events associated with the same request.

tabId

integer. ID of the tab in which the request takes place. Set to -1 if the request isn't related to a tab.

thirdParty

boolean. Indicates whether the request and its content window hierarchy are third party.

timeStamp

number. The time when this event fired, in milliseconds since the epoch.

type

webRequest.ResourceType. The type of resource being requested: for example, "image", "script", "stylesheet".

url

string. Target of the request.

urlClassification

object. The type of tracking associated with the request, if the request is classified by Firefox Tracking Protection. This is an object with these properties:

firstParty: array of strings. Classification flags for the request's first party.
thirdParty: array of strings. Classification flags for the request or its window hierarchy's third parties.

The classification flags include:

fingerprinting and fingerprinting_content: indicates the request is involved in fingerprinting ("an origin found to fingerprint").
- fingerprinting indicates the domain is in the fingerprinting and tracking category. Examples of this type of domain include advertisers who want to associate a profile with the visiting user.
- fingerprinting_content indicates the domain is in the fingerprinting category but not the tracking category. Examples of this type of domain include payment providers who use fingerprinting techniques to identify the visiting user for anti-fraud purposes.
cryptomining and cryptomining_content: similar to the fingerprinting category but for cryptomining resources.
tracking, tracking_ad, tracking_analytics, tracking_social, and tracking_content: indicates the request is involved in tracking. tracking is any generic tracking request, the ad, analytics, social, and content suffixes identify the type of tracker.
emailtracking and emailtracking_content: indicates the request is involved in tracking emails.
any_basic_tracking: a meta flag that combines tracking and fingerprinting flags, excluding tracking_content and fingerprinting_content.
any_strict_tracking: a meta flag that combines all tracking and fingerprinting flags.
any_social_tracking: a meta flag that combines all social tracking flags.

You can find more information on tracker types on the disconnect.me website. The content suffix indicates trackers that track and serve content. Blocking them protects users but can lead to sites breaking or elements not being displayed.

Browser compatibility

BCD tables only load in the browser

Examples

This code changes the "User-Agent" header so the browser identifies itself as Opera 12.16, but only when visiting pages under https://httpbin.org/.

"use strict";

/*
This is the page for which we want to rewrite the User-Agent header.
*/
const targetPage = "https://httpbin.org/*";

/*
Set UA string to Opera 12
*/
const ua =
  "Opera/9.80 (X11; Linux i686; Ubuntu/14.10) Presto/2.12.388 Version/12.16";

/*
Rewrite the User-Agent header to "ua".
*/
function rewriteUserAgentHeader(e) {
  for (const header of e.requestHeaders) {
    if (header.name.toLowerCase() === "user-agent") {
      header.value = ua;
    }
  }
  return { requestHeaders: e.requestHeaders };
}

/*
Add rewriteUserAgentHeader as a listener to onBeforeSendHeaders,
only for the target page.

Make it "blocking" so we can modify the headers.
*/
browser.webRequest.onBeforeSendHeaders.addListener(
  rewriteUserAgentHeader,
  { urls: [targetPage] },
  ["blocking", "requestHeaders"],
);

This code is exactly like the previous example, except that the listener is asynchronous, returning a Promise which is resolved with the new headers:

"use strict";

/*
This is the page for which we want to rewrite the User-Agent header.
*/
const targetPage = "https://httpbin.org/*";

/*
Set UA string to Opera 12
*/
const ua =
  "Opera/9.80 (X11; Linux i686; Ubuntu/14.10) Presto/2.12.388 Version/12.16";

/*
Rewrite the User-Agent header to "ua".
*/
function rewriteUserAgentHeaderAsync(e) {
  const asyncRewrite = new Promise((resolve, reject) => {
    setTimeout(() => {
      for (const header of e.requestHeaders) {
        if (header.name.toLowerCase() === "user-agent") {
          header.value = ua;
        }
      }
      resolve({ requestHeaders: e.requestHeaders });
    }, 2000);
  });

  return asyncRewrite;
}

/*
Add rewriteUserAgentHeader as a listener to onBeforeSendHeaders,
only for the target page.

Make it "blocking" so we can modify the headers.
*/
browser.webRequest.onBeforeSendHeaders.addListener(
  rewriteUserAgentHeaderAsync,
  { urls: [targetPage] },
  ["blocking", "requestHeaders"],
);

Example extensions

user-agent-rewriter

Note: This API is based on Chromium's chrome.webRequest API. This documentation is derived from web_request.json in the Chromium code.