WebUI Explainer

What is “WebUI”?

“WebUI” is a term used to loosely describe parts of Chrome's UI implemented with web technologies (i.e. HTML, CSS, JavaScript).

Examples of WebUI in Chromium:

This document explains how WebUI works.

What's different from a web page?

WebUIs are granted super powers so that they can manage Chrome itself. For example, it'd be very hard to implement the Settings UI without access to many different privacy and security sensitive services. Access to these services are not granted by default.

Only special URLs are granted WebUI “bindings” via the child security process.

Specifically, these bindings:

  • give a renderer access to load chrome: URLS
    • this is helpful for shared libraries, i.e. chrome://resources/
  • allow the browser to execute arbitrary JavaScript in that renderer via CallJavascriptFunction()
  • allow communicating from the renderer to the browser with chrome.send() and friends
  • ignore content settings regarding showing images or executing JavaScript

How chrome: URLs work

A chrome: URL loads a file from disk, memory, or can respond dynamically.

Because Chrome UIs generally need access to the browser (not just the current tab), much of the C++ that handles requests or takes actions lives in the browser process. The browser has many more privileges than a renderer (which is sandboxed and doesn't have file access), so access is only granted for certain URLs.

chrome: protocol

Chrome recognizes a list of special protocols, which it registers while starting up.


  • devtools:
  • chrome-extensions:
  • chrome:
  • file:
  • view-source:

This document mainly cares about the chrome: protocol, but others can also be granted WebUI bindings or have special properties.

chrome: hosts

After registering the chrome: protocol, a set of factories are created. These factories contain a list of valid host names. A valid hostname generates a controller.

In the case of chrome: URLs, these factories are registered early in the browser process lifecycle. Before the first WebUIConfig is registered, the WebUIConfigMap instance is created. This map creates and registers a factory (WebUIConfigMapWebUIControllerFactory) in its constructor. This factory looks at the global WebUIConfigMap, which maps hosts to WebUIConfigs, to see if any of the configs handle the requested URL. It calls the method on the config to create the corresponding controller if it finds a config to handle the URL.

// ChromeBrowserMainParts::PreMainMessageLoopRunImpl():

// Legacy WebUIControllerFactory registration

// Factory for all WebUIs using WebUIConfig will be created here.

When a URL is requested, a new renderer is created to load the URL, and a corresponding class in the browser is set up to handle messages from the renderer to the browser (a RenderFrameHost).

auto* config = config_map_->GetConfig(browser_context, url);
if (!config)
  return nullptr;  // Not a known host; no special access.

return config->CreateWebUIController(web_ui, url);

Configs can be registered with the map by calling map.AddWebUIConfig() in chrome_web_ui_configs.cc:


If a factory knows how to handle a host (returns a WebUIFactoryFunction), the navigation machinery grants the renderer process WebUI bindings via the child security policy.

// RenderFrameHostImpl::AllowBindings():
if (bindings_flags & BINDINGS_POLICY_WEB_UI) {

The factory creates a WebUIController for a tab using the WebUIConfig.

Here's an example:

// Config for chrome://donuts
    : WebUIConfig(content::kChromeUIScheme, chrome::kChromeUIDonutsHost) {}

DonutsUIConfig::~DonutsUIConfig() = default;

DonutsUIConfig::CreateWebUIController(content::WebUI* web_ui,
                                      const GURL& url) {
  return std::make_unique<DonutsUI>(web_ui);

// Controller for chrome://donuts.
class DonutsUI : public content::WebUIController {
  DonutsUI(content::WebUI* web_ui) : content::WebUIController(web_ui) {
    content::WebUIDataSource* source =
            "donuts");  // "donuts" == hostname
    source->AddString("mmmDonuts", "Mmm, donuts!");  // Translations.
    source->AddResourcePath("", IDR_DONUTS_HTML);  // Home page.

    // Handles messages from JavaScript to C++ via chrome.send().

If we assume the contents of IDR_DONUTS_HTML yields:


Visiting chrome://donuts should show in something like:

Delicious success.

By default $i18n{} escapes strings for HTML. $i18nRaw{} can be used for translations that embed HTML, and $i18nPolymer{} can be used for Polymer bindings. See this comment for more information.

C++ classes


WebUI is a high-level class and pretty much all HTML-based Chrome UIs have one. WebUI lives in the browser process, and is owned by a RenderFrameHost. WebUIs have a concrete implementation (WebUIImpl) in content/ and are created in response to navigation events.

A WebUI knows very little about the page it's showing, and it owns a WebUIController that is set after creation based on the hostname of a requested URL.

A WebUI can handle messages itself, but often defers these duties to separate WebUIMessageHandlers, which are generally designed for handling messages on certain topics.

A WebUI can be created speculatively, and are generally fairly lightweight. Heavier duty stuff like hard initialization logic or accessing services that may have side effects are more commonly done in a WebUIController or WebUIMessageHandlers.

WebUI are created synchronously on the UI thread in response to a URL request, and are re-used where possible between navigations (i.e. refreshing a page). Because they run in a separate process and can exist before a corresponding renderer process has been created, special care is required to communicate with the renderer if reliable message passing is required.


A WebUIConfig contains minimal possible logic and information for determining whether a certain subclass of WebUIController should be created for a given URL.

A WebUIConfig holds information about the host and scheme (chrome:// or chrome-untrusted://) that the controller serves.

A WebUIConfig may contain logic to check if the WebUI is enabled for a given BrowserContext and url (e.g., if relevant feature flags are enabled/disabled, if the url path is valid, etc).

A WebUIConfig can invoke the WebUIController's constructor in its CreateWebUIControllerForURL method.

WebUIConfigs are created at startup when factories are registered, so should be lightweight.


A WebUIController is the brains of the operation, and is responsible for application-specific logic, setting up translations and resources, creating message handlers, and potentially responding to requests dynamically. In complex pages, logic is often split across multiple WebUIMessageHandlers instead of solely in the controller for organizational benefits.

A WebUIController is owned by a WebUI, and is created and set on an existing WebUI when the corresponding WebUIConfig is found in the map matching the URL, or when the correct controller is determined via URL inspection in ChromeWebUIControllerFactory. (i.e. chrome://settings creates a generic WebUI with a settings-specific WebUIController).


The WebUIDataSource class provides a place for data to live for WebUI pages.

Examples types of data stored in this class are:

  • static resources (i.e. .html files packed into bundles and pulled off of disk)
  • translations
  • dynamic feature values (i.e. whether a feature is enabled)

Data sources are set up in the browser process (in C++) and are accessed by loading URLs from the renderer.

Below is an example of a simple data source (in this case, Chrome's history page):

content::WebUIDataSource* source = content::WebUIDataSource::CreateAndAdd(
    Profile::FromWebUI(web_ui), "history");

source->AddResourcePath("sign_in_promo.svg", IDR_HISTORY_SIGN_IN_PROMO_SVG);
source->AddResourcePath("synced_tabs.html", IDR_HISTORY_SYNCED_TABS_HTML);

source->AddString("title", IDS_HISTORY_TITLE);
source->AddString("moreFromThisSite", IDS_HISTORY_MORE_FROM_THIS_SITE);


    source, base::make_span(kHistoryResources, kHistoryResourcesSize),
    kGeneratedPath, IDR_HISTORY_HISTORY_HTML);

For more about each of the methods called on WebUIDataSource and the utility method that performs additional configuration, see DataSources and WebUIDataSourceUtils


Because some pages have many messages or share code that sends messages, message handling is often split into discrete classes called WebUIMessageHandlers. These handlers respond to specific invocations from JavaScript.

So, the given C++ code:

void OvenHandler::RegisterMessages() {

void OvenHandler::HandleBakeDonuts(const base::Value::List& args) {

  // IMPORTANT: Fully validate `args`.
  CHECK_EQ(1u, args.size());
  int num_donuts = args[0].GetInt();
  CHECK_GT(num_donuts, 0);

Can be triggered in JavaScript with this example code:

$('bakeDonutsButton').onclick = function() {
  chrome.send('bakeDonuts', [5]);  // bake 5 donuts!

Data Sources


This is a factory method required to create and add a WebUIDataSource. The first argument to Create() is the browser context. The second argument is typically the host name of the page. The caller does not own the result.

Additionally, calling CreateAndAdd() will overwrite any existing data source with the same name.


Using an int reference to a grit string (starts with “IDS” and lives in a .grd or .grdp file), adding a string with a key name will be possible to reference via the $i18n{} syntax (and will be replaced when requested) or later dynamically in JavaScript via loadTimeData.getString() (or getStringF).


Many Web UI data sources need to be set up with a large number of localized strings. Instead of repeatedly calling AddLocalizedString(), create an array of all the strings and use AddLocalizedStrings():

  static constexpr webui::LocalizedString kStrings[] = {
      // Localized strings (alphabetical order).
      {"actionMenuDescription", IDS_HISTORY_ACTION_MENU_DESCRIPTION},
      {"ariaRoleDescription", IDS_HISTORY_ARIA_ROLE_DESCRIPTION},
      {"bookmarked", IDS_HISTORY_ENTRY_BOOKMARKED},


Using an int reference to a grit resource (starts with “IDR” and lives in a .grd or .grdp file), adds a resource to the UI with the specified path.

It's generally a good idea to call AddResourcePath() with the empty path and a resource ID that should be served as the “catch all” resource to respond with. This resource will be served for requests like “chrome://history”, or “chrome://history/pathThatDoesNotExist”. It will not be served for requests that look like they are attempting to fetch a specific file, like “chrome://history/file_that_does_not_exist.js”. This is so that if a user enters a typo when trying to load a subpage like “chrome://history/syncedTabs” they will be redirected to the main history page, instead of seeing an error, but incorrect imports in the source code will fail, so that they can be more easily found and corrected.


Similar to the localized strings, many Web UIs need to add a large number of resource paths. In this case, use AddResourcePaths() to replace repeated calls to AddResourcePath().

  static constexpr webui::ResourcePath kResources[] = {
      {"browser_api.js", IDR_BROWSER_API_JS},
      {"constants.js", IDR_CONSTANTS_JS},
      {"controller.js", IDR_CONTROLLER_JS},

The same method can be leveraged for cases that directly use constants defined by autogenerated grit resources map header files. For example, the autogenerated print_preview_resources_map.h header defines a webui::ResourcePath array named kPrintPreviewResources and a size_t kPrintPreviewResourcesSize. All the resources in this resource map can be added as follows:

      base::make_span(kPrintPreviewResources, kPrintPreviewResourcesSize));


Often a page needs to know whether a feature is enabled. This is a good use case for WebUIDataSource::AddBoolean(). Then, in the Javascript, one can write code like this:

if (loadTimeData.getBoolean('myFeatureIsEnabled')) {

If you really want or need to use AddBoolean() for a dynamic value, make sure to call WebUIDataSource::Update() when the value changes.

WebUI utils for working with data sources

chrome/browser/ui/webui/webui_util.* contains a number of methods to simplify common configuration tasks.


This method performs common configuration tasks on a data source for a Web UI that uses JS modules. When creating a Web UI that uses JS modules, use this utility instead of duplicating the configuration steps it performs elsewhere. Specific setup steps include:

  • Setting the content security policy to allow the data source to load only resources from its own host (e.g. chrome://history), chrome://resources, and chrome://webui-test (used to serve test files).
  • Enabling i18n template replacements by calling UseStringsJs() and EnableReplaceI18nInJS() on the data source.
  • Adding the test loader files to the data source, so that test files can be loaded as JS modules.
  • Setting the resource to load for the empty path.
  • Adding all resources from a GritResourceMap.

Browser (C++) and Renderer (JS) communication


Mojo is used for IPC throughout Chromium, and should generally be used for new WebUIs to communicate between the browser (C++) and the renderer (JS/TS). To use Mojo, you will need to:

  • Write an interface definition for the JS/C++ interface in a mojom file
  • Add a build target in the BUILD.gn file to autogenerate C++ and TypeScript code (“bindings”).
  • Bind the interface on the C++ side and implement any methods to send or receive information from TypeScript.
  • Add the TypeScript bindings file to your WebUI's ts_library() and use them in your TypeScript code.

Mojo Interface Definition

Mojo interfaces are declared in mojom files. For WebUIs, these normally live alongside the C++ code in chrome/browser/ui/webui. For example:


module donuts.mojom;

// Factory ensures that the Page and PageHandler interfaces are always created
// together without requiring an initialization call from the WebUI to the
// handler.
interface PageHandlerFactory {
  CreatePageHandler(pending_remote<Page> page,
                    pending_receiver<PageHandler> handler);

// Called from TS side of chrome://donuts (Renderer -> Browser)
interface PageHandler {

  BakeDonuts(uint32 num_donuts);

  // Expects a response from the browser.
  GetNumberOfDonuts() => (uint32 num_donuts);

// Called from C++ side of chrome://donuts. (Browser -> Renderer)
interface Page {
  DonutsBaked(uint32 num_donuts);

BUILD.gn mojo target

mojom() is the build rule used to generate mojo bindings. It can be set up as follows:



mojom("mojo_bindings") {
  sources = [ "donuts.mojom" ]
  webui_module_path = "/"

Setting up C++ bindings

The WebUIController class should inherit from ui::MojoWebUIController and from the PageHandlerFactory class defined in the mojom file.


class DonutsPageHandler;

class DonutsUI : public ui::MojoWebUIController,
                 public donuts::mojom::PageHandlerFactory {
  explicit DonutsUI(content::WebUI* web_ui);

  DonutsUI(const DonutsUI&) = delete;
  DonutsUI& operator=(const DonutsUI&) = delete;

  ~DonutsUI() override;

  // Instantiates the implementor of the mojom::PageHandlerFactory mojo
  // interface passing the pending receiver that will be internally bound.
  void BindInterface(
      mojo::PendingReceiver<donuts::mojom::PageHandlerFactory> receiver);

  // donuts::mojom::PageHandlerFactory:
  void CreatePageHandler(
      mojo::PendingRemote<donuts::mojom::Page> page,
      mojo::PendingReceiver<donuts::mojom::PageHandler> receiver) override;

  std::unique_ptr<DonutsPageHandler> page_handler_;

  mojo::Receiver<donuts::mojom::PageHandlerFactory> page_factory_receiver_{



DonutsUI::DonutsUI(content::WebUI* web_ui)
    : ui::MojoWebUIController(web_ui, true) {
  // Normal constructor steps (e.g. setting up data source) go here.


DonutsUI::~DonutsUI() = default;

void DonutsUI::BindInterface(
    mojo::PendingReceiver<donuts::mojom::PageHandlerFactory> receiver) {

void DonutsUI::CreatePageHandler(
    mojo::PendingRemote<donuts::mojom::Page> page,
    mojo::PendingReceiver<donuts::mojom::PageHandler> receiver) {
  page_handler_ = std::make_unique<DonutsPageHandler>(
      std::move(receiver), std::move(page));

You also need to register the PageHandlerFactory to your controller in chrome/browser/chrome_browser_interface_binders.cc:


Using C++ bindings for communication

The WebUI message handler should inherit from the Mojo PageHandler class.


#include "chrome/browser/ui/webui/donuts/donuts.mojom.h"
#include "mojo/public/cpp/bindings/receiver.h"
#include "mojo/public/cpp/bindings/remote.h"

class DonutsPageHandler : public donuts::mojom::PageHandler {
      mojo::PendingReceiver<donuts::mojom::PageHandler> receiver,
      mojo::PendingRemote<donuts::mojom::Page> page);

  DonutsPageHandler(const DonutsPageHandler&) = delete;
  DonutsPageHandler& operator=(const DonutsPageHandler&) = delete;

  ~DonutsPageHandler() override;

  // Triggered by some outside event
  void DonutsPageHandler::OnBakingDonutsFinished(uint32_t num_donuts);

  // donuts::mojom::PageHandler:
  void StartPilotLight() override;
  void BakeDonuts(uint32_t num_donuts) override;
  void GetNumberOfDonuts(GetNumberOfDonutsCallback callback) override;

The message handler needs to implement all the methods on the PageHandler interface.


    mojo::PendingReceiver<donuts::mojom::PageHandler> receiver,
    mojo::PendingRemote<donuts::mojom::Page> page)
    : receiver_(this, std::move(receiver)),
      page_(std::move(page)) {

DonutsPageHandler::~DonutsPageHandler() {

// Triggered by outside asynchronous event; sends information to the renderer.
void DonutsPageHandler::OnBakingDonutsFinished(uint32_t num_donuts) {

// Triggered by startPilotLight() call in TS.
void DonutsPageHandler::StartPilotLight() {

// Triggered by bakeDonuts() call in TS.
void DonutsPageHandler::BakeDonuts(int32_t num_donuts) {

// Triggered by getNumberOfDonuts() call in TS; sends a response back to the
// renderer.
void DonutsPageHandler::GetNumberOfDonuts(GetNumberOfDonutsCallback callback) {
  uint32_t result = GetOven()->GetNumberOfDonuts();

Setting Up TypeScript bindings

For WebUIs using the build_webui() rule, the TypeScript mojo bindings can be added to the build and served from the root (e.g. chrome://donuts/donuts.mojom-webui.js) by adding the following arguments to build_webui():


build_webui("build") {
  # ... Other arguments go here
  mojo_files_deps =
      [ "//chrome/browser/ui/webui/donuts:mojo_bindings_ts__generator" ]
  mojo_files = [
  # ... Other arguments can go here

It is often helpful to wrap the TypeScript side of Mojo setup in a BrowserProxy class:


import {PageCallbackRouter, PageHandlerFactory, PageHandlerInterface, PageHandlerRemote} from './donuts.mojom-webui.js';

class BrowserProxy {
  callbackRouter: PageCallbackRouter;
  handler: PageHandlerInterface;

  constructor() {
    this.callbackRouter = new PageCallbackRouter();

    this.handler = new PageHandlerRemote();

    const factory = PageHandlerFactory.getRemote();
        (this.handler as PageHandlerRemote).$.bindNewPipeAndPassReceiver());

  static getInstance(): BrowserProxy {
    return instance || (instance = new BrowserProxy());

  static setInstance(obj: BrowserProxy) {
    instance = obj;

let instance: BrowserProxy|null = null;

Using TypeScript bindings for communication

The callbackRouter (PageCallbackRouter) can be used to add listeners for asynchronous events sent from the browser.

The handler (PageHandlerRemote) can be used to send messages from the renderer to the browser. For interface methods that require a browser response, calling the method returns a promise. The promise will be resolved with the response from the browser.


import {BrowserProxy} from './browser_proxy.js';

let numDonutsBaked: number = 0;

window.onload = function() {
  // Other page initialization steps go here
  const proxy = BrowserProxy.getInstance();
  // Tells the browser to start the pilot light.
  // Adds a listener for the asynchronous "donutsBaked" event.
    (numDonuts: number) => {
      numDonutsBaked += numDonuts;

function CheckNumberOfDonuts() {
  // Requests the number of donuts from the browser, and alerts with the
  // response.
      (numDonuts: number) => {
        alert('Yay, there are ' + numDonuts + ' delicious donuts left!');

function BakeDonuts(numDonuts: number) {
  // Tells the browser to bake |numDonuts| donuts.

Pre-Mojo alternative: chrome.send()/WebUIMessageHandler

Most Chrome WebUIs were added before the introduction of Mojo, and use the older style WebUIMessageHandler + chrome.send() pattern. The following sections detail the methods in WebUIMessageHandler and the corresponding communication methods in TypeScript/JavaScript and how to use them.


A tab that has been used for settings UI may be reloaded, or may navigate to an external origin. In both cases, one does not want callbacks from C++ to Javascript to run. In the former case, the callbacks will occur when the Javascript doesn't expect them. In the latter case, sensitive information may be delivered to an untrusted origin.

Therefore each message handler maintains a boolean that describes whether delivering callbacks to Javascript is currently appropriate. This boolean is set by calling AllowJavascript, which should be done when handling a call from Javascript, because that indicates that the page is ready for the subsequent callback. (See design doc.) If the tab navigates or reloads, DisallowJavascript is called to clear the flag.

Therefore, before each callback from C++ to Javascript, the flag must be tested by calling IsJavascriptAllowed. If false, then the callback must be dropped. (When the flag is false, calling ResolveJavascriptCallback will crash. See design doc.)

Also beware of ABA issues: Consider the case where an asynchronous operation is started, the settings page is reloaded, and the user triggers another operation using the original message handler. The javascript_allowed_ boolean will be true, but the original callback should still be dropped because it relates to a operation that was discarded by the reload. (Reloading settings UI does not cause message handler objects to be deleted.)

Thus a message handler may override OnJavascriptDisallowed to learn when pending callbacks should be canceled.

In the JS:

window.onload = function() {

In the C++:

void OvenHandler::HandleStartPilotLight(const base::Value::List& /*args*/) {
  // CallJavascriptFunction() and FireWebUIListener() are now safe to do.


When the browser process needs to tell the renderer/JS of an event or otherwise execute code, it can use CallJavascriptFunction().

void OvenHandler::OnPilotLightExtinguished() {

This works by crafting a string to be evaluated in the renderer. Any arguments to the call are serialized to JSON and the parameter list is wrapped with

// See WebUI::GetJavascriptCall() for specifics:
"functionCallName(" + argumentsAsJson + ")"

and sent to the renderer via a FrameMsg_JavaScriptExecuteRequest IPC message.

While this works, it implies that:

  • a global method must exist to successfully run the Javascript request
  • any method can be called with any parameter (far more access than required in practice)

^ These factors have resulted in less use of CallJavascriptFunction() in the webui codebase. This functionality can easily be accomplished with the following alternatives:


FireWebUIListener() is used to notify a registered set of listeners that an event has occurred. This is generally used for events that are not guaranteed to happen in timely manner, or may be caused to happen by unpredictable events (i.e. user actions).

Here‘s some example to detect a change to Chrome’s theme:

addWebUiListener("theme-changed", refreshThemeStyles);

This Javascript event listener can be triggered in C++ via:

void MyHandler::OnThemeChanged() {

Because it‘s not clear when a user might want to change their theme nor what theme they’ll choose, this is a good candidate for an event listener.

If you simply need to get a response in Javascript from C++, consider using sendWithPromise() and ResolveJavascriptCallback.


OnJavascriptDisallowed() is a lifecycle method called in response to AllowJavascript(). It is a good place to register observers of global services or other callbacks that might call at unpredictable times.

For example:

class MyHandler : public content::WebUIMessageHandler {
  MyHandler() {
    GetGlobalService()->AddObserver(this);  // <-- DON'T DO THIS.
  void OnGlobalServiceEvent() {

Because browser-side C++ handlers are created before a renderer is ready, the above code may result in calling FireWebUIListener before the renderer is ready, which may result in dropped updates or accidentally running Javascript in a renderer that has navigated to a new URL.

A safer way to set up communication is:

class MyHandler : public content::WebUIMessageHandler {
  void OnJavascriptAllowed() override {
    observation_.Observe(GetGlobalService());  // <-- DO THIS.
  void OnJavascriptDisallowed() override {
    observation_.Reset();  // <-- AND THIS.
  base::ScopedObservation<MyHandler, GlobalService> observation_{this};  // <-- ALSO HANDY.

when a renderer has been created and the document has loaded enough to signal to the C++ that it's ready to respond to messages.


OnJavascriptDisallowed is a lifecycle method called when it‘s unclear whether it’s safe to send JavaScript messsages to the renderer.

There's a number of situations that result in this method being called:

  • renderer doesn't exist yet
  • renderer exists but isn't ready
  • renderer is ready but application-specific JS isn't ready yet
  • tab refresh
  • renderer crash

Though it‘s possible to programmatically disable Javascript, it’s uncommon to need to do so.

Because there‘s no single strategy that works for all cases of a renderer’s state (i.e. queueing vs dropping messages), these lifecycle methods were introduced so a WebUI application can implement these decisions itself.

Often, it makes sense to disconnect from observers in OnJavascriptDisallowed():

void OvenHandler::OnJavascriptDisallowed() {

Because OnJavascriptDisallowed() is not guaranteed to be called before a WebUIMessageHandler's destructor, it is often advisable to use some form of scoped observer that automatically unsubscribes on destruction but can also imperatively unsubscribe in OnJavascriptDisallowed().


This method is called in response to sendWithPromise() to reject the issued Promise. This runs the rejection (second) callback in the Promise's executor and any catch() callbacks in the chain.

void OvenHandler::HandleBakeDonuts(const base::Value::List& args) {
  if (!GetOven()->HasGas()) {
                             base::StringValue("need gas to cook the donuts!"));

This method is basically just a CallJavascriptFunction() wrapper that calls a global “cr.webUIResponse” method with a success value of false.

// WebUIMessageHandler::RejectJavascriptCallback():
CallJavascriptFunction("cr.webUIResponse", callback_id, base::Value(false),

See also: ResolveJavascriptCallback


This method is called in response to sendWithPromise() to fulfill an issued Promise, often with a value. This results in runnings any fulfillment (first) callbacks in the associate Promise executor and any registered then() callbacks.

So, given this TypeScript code:

sendWithPromise('bakeDonuts', [5]).then(function(numDonutsBaked: number) {
  shop.donuts += numDonutsBaked;

Some handling C++ might do this:

void OvenHandler::HandleBakeDonuts(const base::Value::List& args) {
  double num_donuts_baked = GetOven()->BakeDonuts();
  ResolveJavascriptCallback(args[0], base::Value(num_donuts_baked));


When the JavaScript window object is created, a renderer is checked for WebUI bindings.

// RenderFrameImpl::DidClearWindowObject():
if (enabled_bindings_ & BINDINGS_POLICY_WEB_UI)

If the bindings exist, a global chrome.send() function is exposed to the renderer:

// WebUIExtension::Install():
v8::Local<v8::Object> chrome = GetOrCreateChromeObject(isolate, context);
chrome->Set(gin::StringToSymbol(isolate, "send"),

The chrome.send() method takes a message name and argument list.

chrome.send('messageName', [arg1, arg2, ...]);

The message name and argument list are serialized to JSON and sent via the FrameHostMsg_WebUISend IPC message from the renderer to the browser.

// In the renderer (WebUIExtension::Send()):
render_frame->Send(new FrameHostMsg_WebUISend(render_frame->GetRoutingID(),
                                              message, *content));
// In the browser (WebUIImpl::OnMessageReceived()):

The browser-side code does a map lookup for the message name and calls the found callback with the deserialized arguments:

// WebUIImpl::ProcessWebUIMessage():


WebUI listeners are a convenient way for C++ to inform JavaScript of events.

Older WebUI code exposed public methods for event notification, similar to how responses to chrome.send() used to work. They both resulted in global namespace pollution, but it was additionally hard to stop listening for events in some cases. addWebUiListener is preferred in new code.

Adding WebUI listeners creates and inserts a unique ID into a map in JavaScript, just like sendWithPromise().

addWebUiListener can be imported from ‘chrome://resources/js/cr.js’.

// addWebUiListener():
webUIListenerMap[eventName] = webUIListenerMap[eventName] || {};
webUIListenerMap[eventName][createUid()] = callback;

The C++ responds to a globally exposed function (cr.webUIListenerCallback) with an event name and a variable number of arguments.

// WebUIMessageHandler:
template <typename... Values>
void FireWebUIListener(const std::string& event_name, const Values&... values) {
  CallJavascriptFunction("cr.webUIListenerCallback", base::Value(event_name),

C++ handlers call this FireWebUIListener method when an event occurs that should be communicated to the JavaScript running in a tab.

void OvenHandler::OnBakingDonutsFinished(size_t num_donuts) {
  FireWebUIListener("donuts-baked", base::Value(num_donuts));

TypeScript can listen for WebUI events via:

let donutsReady: number = 0;
addWebUiListener('donuts-baked', function(numFreshlyBakedDonuts: number) {
  donutsReady += numFreshlyBakedDonuts;


sendWithPromise() is a wrapper around chrome.send(). It's used when triggering a message requires a response:

chrome.send('getNumberOfDonuts');  // No easy way to get response!

In older WebUI pages, global methods were exposed simply so responses could be sent. This is discouraged as it pollutes the global namespace and is harder to make request specific or do from deeply nested code.

In newer WebUI pages, you see code like this:

sendWithPromise('getNumberOfDonuts').then(function(numDonuts: number) {
  alert('Yay, there are ' + numDonuts + ' delicious donuts left!');

Note that sendWithPromise can be imported from ‘chrome://resources/js/cr.js’;

On the C++ side, the message registration is similar to chrome.send() except that the first argument in the message handler's list is a callback ID. That ID is passed to ResolveJavascriptCallback(), which ends up resolving the Promise in JavaScript/TypeScript and calling the then() function.

void DonutHandler::HandleGetNumberOfDonuts(const base::Value::List& args) {

  const base::Value& callback_id = args[0];
  size_t num_donuts = GetOven()->GetNumberOfDonuts();
  ResolveJavascriptCallback(callback_id, base::Value(num_donuts));

Under the covers, a map of Promises are kept in JavaScript.

The callback ID is just a namespaced, ever-increasing number. It's used to insert a Promise into the JS-side map when created.

// sendWithPromise():
var id = methodName + '_' + uidCounter++;
chromeSendResolverMap[id] = new PromiseResolver;
chrome.send(methodName, [id].concat(args));

The corresponding number is used to look up a Promise and reject or resolve it when the outcome is known.

// cr.webUIResponse():
var resolver = chromeSendResolverMap[id];
if (success)

This approach still relies on the C++ calling a globally exposed method, but reduces the surface to only a single global (cr.webUIResponse) instead of many. It also makes per-request responses easier, which is helpful when multiple are in flight.

Security considerations

Because WebUI pages are highly privileged, they are often targets for attack, since taking control of a WebUI page can sometimes be sufficient to escape Chrome's sandbox. To make sure that the special powers granted to WebUI pages are safe, WebUI pages are restricted in what they can do:

  • WebUI pages cannot embed http/https resources
  • WebUI pages cannot issue http/https fetches

In the rare case that a WebUI page really needs to include web content, the safe way to do this is by using an <iframe> tag. Chrome's security model gives process isolation between the WebUI and the web content. However, some extra precautions need to be taken, because there are properties of the page that are accessible cross-origin and malicious code can take advantage of such data to attack the WebUI. Here are some things to keep in mind:

  • The WebUI page can receive postMessage payloads from the web and should ensure it verifies any messages as they are not trustworthy.
  • The entire frame tree is visible to the embedded web content, including ancestor origins.
  • The web content runs in the same StoragePartition and Profile as the WebUI, which reflect where the WebUI page was loaded (e.g., the default profile, Incognito, etc). The corresponding user credentials will thus be available to the web content inside the WebUI, possibly showing the user as signed in.

Note: WebUIs have a default Content Security Policy which disallows embedding any frames. If you want to include any web content in an you will need to update the policy for your WebUI. When doing so, allow only known origins and avoid making the policy more permissive than strictly necessary.

Alternatively, a <webview> tag can be used, which runs in a separate StoragePartition, a separate frame tree, and restricts postMessage communication by default. Note that <webview> is only available on desktop platforms.

JavaScript Error Reporting

By default, errors in the JavaScript or TypeScript of a WebUI page will generate error reports which appear in Google's internal go/crash reports page. These error reports will only be generated for Google Chrome builds, not Chromium or other Chromium-based browsers.

Specifically, an error report will be generated when the JavaScript or TypeScript for a WebUI-based chrome:// page does one of the following:

  • Generates an uncaught exception,
  • Has a promise which is rejected, and no rejection handler is provided, or
  • Calls console.error().

Such errors will appear alongside other crashes in the product_name=Chrome_ChromeOS, product_name=Chrome_Lacros, or product_name=Chrome_Linux lists on go/crash.

The signature of the error is the error message followed by the URL on which the error appeared. For example, if chrome://settings/lazy_load.js throws a TypeError with a message Cannot read properties of null (reading 'select') and does not catch it, the magic signature would be

Uncaught TypeError: Cannot read properties of null (reading 'select') (chrome://settings)

To avoid spamming the system, only one error report with a given message will be generated per hour.

If you are getting error reports for an expected condition, you can turn off the reports simply by changing console.error() into console.warn(). For instance, if JavaScript is calling console.error() when the user tries to connect to an unavailable WiFi network at the same time the page shows the user an error message, the console.error() should be replaced with a console.warn().

If you wish to get more control of the JavaScript error messages, for example to change the product name or to add additional data, you may wish to switch to using CrashReportPrivate.reportError(). If you do so, be sure to override WebUIController::IsJavascriptErrorReportingEnabled() to return false for your page; this will avoid generating redundant error reports.

Are JavaScript errors actually crashes?

JavaScript errors are not “crashes” in the C++ sense. They do not stop a process from running, they do not cause a “sad tab” page. Some tooling refers to them as crashes because they are going through the same pipeline as the C++ crashes, and that pipeline was originally designed to handle crashes.

How much impact does this JavaScript error have?

That depends on the JavaScript error. In some cases, the errors have no user impact; for instance, the “unavailable WiFi network calling console.error()” example above. In other cases, JavaScript errors may be serious errors that block the user from completing critical user journeys. For example, if the JavaScript is supposed to un-hide one of several variants of settings page, but the JavaScript has an unhandled exception before un-hiding any of them, then the user will see a blank page and be unable to change that setting.

Because it is difficult to automatically determine the severity of a given error, JavaScript errors are currently all classified as “WARNING” level when computing stability metrics.

Known issues

  1. Error reporting is currently enabled only on ChromeOS (ash and Lacros) and Linux.
  2. Errors are only reported for chrome:// URLs.
  3. Unhandled promise rejections do not have a good stack.
  4. The line numbers and column numbers in the stacks are for the minified JavaScript and do not correspond to the line and column numbers of the original source files.
  5. Error messages with variable strings do not group well. For example, if the error message includes the name of a network, each network name will be its own signature.

See also