1. 11 Worklets
    1. 11.1 Introduction
      1. 11.1.1 Motivations
      2. 11.1.2 Code idempotence
      3. 11.1.3 Speculative evaluation
    2. 11.2 Examples
      1. 11.2.1 Loading scripts
      2. 11.2.2 Registering a class and invoking its methods
    3. 11.3 Infrastructure
      1. 11.3.1 The global scope
        1. Agents and event loops
        2. Creation and termination
        3. Script settings for worklets
      2. 11.3.2 The Worklet class
      3. 11.3.3 The worklet's lifetime

11 Worklets

11.1 Introduction

This section is non-normative.

Worklets are a piece of specification infrastructure which can be used for running scripts independent of the main JavaScript execution environment, while not requiring any particular implementation model.

The worklet infrastructure specified here cannot be used directly by web developers. Instead, other specifications build upon it to create directly-usable worklet types, specialized for running in particular parts of the browser implementation pipeline.

11.1.1 Motivations

This section is non-normative.

Allowing extension points to rendering, or other sensitive parts of the implementation pipeline such as audio output, is difficult. If extension points were done with full access to the APIs available on Window, engines would need to abandon previously-held assumptions for what could happen in the middle of those phases. For example, during the layout phase, rendering engines assume that no DOM will be modified.

Additionally, defining extension points in the Window environment would restrict user agents to performing work in the same thread as the Window object. (Unless implementations added complex, high-overhead infrastructure to allow thread-safe APIs, as well as thread-joining guarantees.)

Worklets are designed to allow extension points, while keeping guarantees that user agents currently rely on. This is done through new global environments, based on subclasses of WorkletGlobalScope.

Worklets are similar to web workers. However, they:

As worklets have relatively high overhead, they are best used sparingly. Due to this, a given WorkletGlobalScope is expected to be shared between multiple separate scripts. (This is similar to how a single Window is shared between multiple separate scripts.)

Worklets are a general technology that serve different use cases. Some worklets, such as those defined in CSS Painting API, provide extension points intended for stateless, idempotent, and short-running computations, which have special considerations as described in the next couple of sections. Others, such as those defined in Web Audio API, are used for stateful, long-running operations. [CSSPAINT] [WEBAUDIO]

11.1.2 Code idempotence

Some specifications which use worklets are intended to allow user agents to parallelize work over multiple threads, or to move work between threads as required. In these specifications, user agents might invoke methods on a web-developer-provided class in an implementation-defined order.

As a result of this, to prevent interoperability issues, authors who register classes on such WorkletGlobalScopes should make their code idempotent. That is, a method or set of methods on the class should produce the same output given a particular input.

This specification uses the following techniques in order to encourage authors to write code in an idempotent way:

Together, these restrictions help prevent two different scripts from sharing state using properties of the global object.

Additionally, specifications which use worklets and intend to allow implementation-defined behavior must obey the following:

11.1.3 Speculative evaluation

Some specifications which use worklets can invoke methods on a web-developer-provided class based on the state of the user agent. To increase concurrency between threads, a user agent may invoke a method speculatively, based on potential future states.

In these specifications, user agents might invoke such methods at any time, and with any arguments, not just ones corresponding to the current state of the user agent. The results of such speculative evaluations are not displayed immediately, but can be cached for use if the user agent state matches the speculated state. This can increase the concurrency between the user agent and worklet threads.

As a result of this, to prevent interoperability risks between user agents, authors who register classes on such WorkletGlobalScopes should make their code stateless. That is, the only effect of invoking a method should be its result, and not any side effects such as updating mutable state.

The same techniques which encourage code idempotence also encourage authors to write stateless code.

11.2 Examples

This section is non-normative.

For these examples, we'll use a fake worklet. The Window object provides two Worklet instances, which each run code in their own collection of FakeWorkletGlobalScopes:

partial interface Window {
  [SameObject, SecureContext] readonly attribute Worklet fakeWorklet1;
  [SameObject, SecureContext] readonly attribute Worklet fakeWorklet2;

Each Window has two Worklet instances, fake worklet 1 and fake worklet 2. Both of these have their worklet global scope type set to FakeWorkletGlobalScope, and their worklet destination type set to "fakeworklet". User agents should create at least two FakeWorkletGlobalScope instances per worklet.

"fakeworklet" is not actually a valid destination per Fetch. But this illustrates how real worklets would generally have their own worklet-type-specific destination. [FETCH]

The fakeWorklet1 getter steps are to return this's fake worklet 1.

The fakeWorklet2 getter steps are to return this's fake worklet 2.

interface FakeWorkletGlobalScope : WorkletGlobalScope {
  undefined registerFake(DOMString type, Function classConstructor);

Each FakeWorkletGlobalScope has a registered class constructors map, which is an ordered map, initially empty.

The registerFake(type, classConstructor) method steps are to set this's registered class constructors map[type] to classConstructor.

11.2.1 Loading scripts

This section is non-normative.

To load scripts into fake worklet 1, a web developer would write:


Note that which script finishes fetching and runs first is dependent on network timing: it could be either script1.mjs or script2.mjs. This generally won't matter for well-written scripts intended to be loaded in worklets, if they follow the suggestions about preparing for speculative evaluation.

If a web developer wants to perform a task only after the scripts have successfully run and loaded into some worklets, they could write:

]).then(() => {
    // Do something which relies on those scripts being loaded.

Another important point about script-loading is that loaded scripts can be run in multiple WorkletGlobalScopes per Worklet, as discussed in the section on code idempotence. In particular, the specification above for fake worklet 1 and fake worklet 2 require this. So, consider a scenario such as the following:

// script.mjs
console.log("Hello from a FakeWorkletGlobalScope!");
// app.mjs

This could result in output such as the following from a user agent's console:

[fakeWorklet1#1] Hello from a FakeWorkletGlobalScope!
[fakeWorklet1#4] Hello from a FakeWorkletGlobalScope!
[fakeWorklet1#2] Hello from a FakeWorkletGlobalScope!
[fakeWorklet1#3] Hello from a FakeWorkletGlobalScope!

If the user agent at some point decided to kill and restart the third instance of FakeWorkletGlobalScope, the console would again print [fakeWorklet1#3] Hello from a FakeWorkletGlobalScope! when this occurs.

11.2.2 Registering a class and invoking its methods

This section is non-normative.

Let's say that one of the intended usages of our fake worklet by web developers is to allow them to customize the highly-complex process of boolean negation. They might register their customization as follows:

// script.mjs
registerFake('negation-processor', class {
  process(arg) {
    return !arg;
// app.mjs

To make use of such registered classes, the specification for fake worklets could define a find the opposite of true algorithm, given a Worklet worklet:

  1. Optionally, create a worklet global scope for worklet.

  2. Let workletGlobalScope be one of worklet's global scopes, chosen in an implementation-defined manner.

  3. Let classConstructor be workletGlobalScope's registered class constructors map["negation-processor"].

  4. Let classInstance be the result of constructing classConstructor, with no arguments.

  5. Let function be Get(classInstance, "process"). Rethrow any exceptions.

  6. Let callback be the result of converting function to a Web IDL Function instance.

  7. Return the result of invoking callback with the arguments « true » and with classInstance as the callback this value.

Another, perhaps better, specification architecture would be to extract the "process" property and convert it into a Function at registration time, as part of the registerFake() method steps.

11.3 Infrastructure

11.3.1 The global scope

Subclasses of WorkletGlobalScope are used to create global objects wherein code loaded into a particular Worklet can execute.

[Exposed=Worklet, SecureContext]
interface WorkletGlobalScope {};

Other specifications are intended to subclass WorkletGlobalScope, adding APIs to register a class, as well as other APIs specific for their worklet type.

Each WorkletGlobalScope has an associated module map. It is a module map, initially empty. Agents and event loops

This section is non-normative.

Each WorkletGlobalScope is contained in its own worklet agent, which has its corresponding event loop. However, in practice, implementation of these agents and event loops is expected to be different from most others.

A worklet agent exists for each WorkletGlobalScope since, in theory, an implementation could use a separate thread for each WorkletGlobalScope instance, and allowing this level of parallelism is best done using agents. However, because their [[CanBlock]] value is false, there is no requirement that agents and threads are one-to-one. This allows implementations the freedom to execute scripts loaded into a worklet on any thread, including one running code from other agents with [[CanBlock]] of false, such as the thread of a similar-origin window agent ("the main thread"). Contrast this with dedicated worker agents, whose true value for [[CanBlock]] effectively requires them to get a dedicated operating system thread.

Worklet event loops are also somewhat special. They are only used for tasks associated with addModule(), tasks wherein the user agent invokes author-defined methods, and microtasks. Thus, even though the event loop processing model specifies that all event loops run continuously, implementations can achieve observably-equivalent results using a simpler strategy, which just invokes author-provided methods and then relies on that process to perform a microtask checkpoint. Creation and termination

To create a worklet global scope for a Worklet worklet:

  1. Let outsideSettings be worklet's relevant settings object.

  2. Let agent be the result of obtaining a worklet agent given outsideSettings. Run the rest of these steps in that agent.

  3. Let realmExecutionContext be the result of creating a new realm given agent and the following customizations:

  4. Let workletGlobalScope be the global object of realmExecutionContext's Realm component.

  5. Let insideSettings be the result of setting up a worklet environment settings object given realmExecutionContext and outsideSettings.

  6. Let pendingAddedModules be a clone of worklet's added modules list.

  7. Let runNextAddedModule be the following steps:

    1. If pendingAddedModules is not empty, then:

      1. Let moduleURL be the result of dequeueing from pendingAddedModules.

      2. Fetch a worklet script graph given moduleURL, insideSettings, worklet's worklet destination type, what credentials mode?, insideSettings, worklet's module responses map, and with the following steps given script:

        This will not actually perform a network request, as it will just reuse responses from worklet's module responses map. The main purpose of this step is to create a new workletGlobalScope-specific module script from the response.

        1. Assert: script is not null, since the fetch succeeded and the source text was successfully parsed when worklet's module responses map was initially populated with moduleURL.

        2. Run a module script given script.

        3. Run runNextAddedModule.

      3. Abort these steps.
    2. Append workletGlobalScope to outsideSettings's global object's associated Document's worklet global scopes.

    3. Append workletGlobalScope to worklet's global scopes.

    4. Run the responsible event loop specified by insideSettings.

  8. Run runNextAddedModule.

To terminate a worklet global scope given a WorkletGlobalScope workletGlobalScope:

  1. Let eventLoop be workletGlobalScope's relevant agent's event loop.

  2. If there are any tasks queued in eventLoop's task queues, discard them without processing them.

  3. Wait for eventLoop to complete the currently running task.

  4. If the previous step doesn't complete within an implementation-defined period of time, then abort the script currently running in the worklet.

  5. Destroy eventLoop.

  6. Remove workletGlobalScope from the global scopes of the Worklet whose global scopes contains workletGlobalScope.

  7. Remove workletGlobalScope from the worklet global scopes of the Document whose worklet global scopes contains workletGlobalScope. Script settings for worklets

To set up a worklet environment settings object, given a JavaScript execution context executionContext and an environment settings object outsideSettings:

  1. Let origin be a unique opaque origin.

  2. Let inheritedAPIBaseURL be outsideSettings's API base URL.

  3. Let inheritedPolicyContainer be a clone of outsideSettings's policy container.

  4. Let realm be the value of executionContext's Realm component.

  5. Let workletGlobalScope be realm's global object.

  6. Let settingsObject be a new environment settings object whose algorithms are defined as follows:

    The realm execution context

    Return executionContext.

    The module map

    Return workletGlobalScope's module map.

    The API base URL

    Return inheritedAPIBaseURL.

    Unlike workers or other globals derived from a single resource, worklets have no primary resource; instead, multiple scripts, each with their own URL, are loaded into the global scope via worklet.addModule(). So this API base URL is rather unlike that of other globals. However, so far this doesn't matter, as no APIs available to worklet code make use of the API base URL.

    The origin

    Return origin.

    The policy container

    Return inheritedPolicyContainer.

    The cross-origin isolated capability

    Return TODO.

    The time origin

    Assert: this algorithm is never called, because the time origin is not available in a worklet context.

  7. Set settingsObject's id to a new unique opaque string, creation URL to inheritedAPIBaseURL, top-level creation URL to null, top-level origin to outsideSettings's top-level origin, target browsing context to null, and active service worker to null.

  8. Set realm's [[HostDefined]] field to settingsObject.

  9. Return settingsObject.

11.3.2 The Worklet class


Support in all current engines.

Edge (Legacy)?Internet ExplorerNo
Firefox Android?Safari iOS?Chrome Android?WebView Android?Samsung Internet?Opera Android?

The Worklet class provides the capability to add module scripts into its associated WorkletGlobalScopes. The user agent can then create classes registered on the WorkletGlobalScopes and invoke their methods.

[Exposed=Window, SecureContext]
interface Worklet {
  [NewObject] Promise<undefined> addModule(USVString moduleURL, optional WorkletOptions options = {});

dictionary WorkletOptions {
  RequestCredentials credentials = "same-origin";

Specifications that create Worklet instances must specify the following for a given instance:

await worklet.addModule(moduleURL[, { credentials }])


Support in all current engines.

Edge (Legacy)?Internet ExplorerNo
Firefox Android?Safari iOS?Chrome Android?WebView Android?Samsung Internet?Opera Android?

Loads and executes the module script given by moduleURL into all of worklet's global scopes. It can also create additional global scopes as part of this process, depending on the worklet type. The returned promise will fulfill once the script has been successfully loaded and run in all global scopes.

The credentials option can be set to a credentials mode to modify the script-fetching process. It defaults to "same-origin".

Any failures in fetching the script or its dependencies will cause the returned promise to be rejected with an "AbortError" DOMException. Any errors in parsing the script or its dependencies will cause the returned promise to be rejected with the exception generated during parsing.

A Worklet has a list of global scopes, which contains instances of the Worklet's worklet global scope type. It is initially empty.

A Worklet has an added modules list, which is a list of URLs, initially empty. Access to this list should be thread-safe.

A Worklet has a module responses map, which is an ordered map from URLs to either "fetching" or tuples consisting of a response and either null, failure, or a byte sequence representing the response body. This map is initially empty, and access to it should be thread-safe.

The added modules list and module responses map exist to ensure that WorkletGlobalScopes created at different times get equivalent module scripts run in them, based on the same source text. This allows the creation of additional WorkletGlobalScopes to be transparent to the author.

In practice, user agents are not expected to implement these data structures, and the algorithms that consult them, using thread-safe programming techniques. Instead, when addModule() is called, user agents can fetch the module graph on the main thread, and send the fetched source text (i.e., the important data contained in the module responses map) to each thread which has a WorkletGlobalScope.

Then, when a user agent creates a new WorkletGlobalScope for a given Worklet, it can simply send the map of fetched source text and the list of entry points from the main thread to the thread containing the new WorkletGlobalScope.

The addModule(moduleURL, options) method steps are:

  1. Let outsideSettings be the relevant settings object of this.

  2. Let moduleURLRecord be the result of encoding-parsing a URL given moduleURL, relative to outsideSettings.

  3. If moduleURLRecord is failure, then return a promise rejected with a "SyntaxError" DOMException.

  4. Let promise be a new promise.

  5. Run the following steps in parallel:

    1. If this's global scopes is empty, then:

      1. Create a worklet global scope given this.

      2. Optionally, create additional global scope instances given this, depending on the specific worklet in question and its specification.

      3. Wait for all steps of the creation process(es) — including those taking place within the worklet agents — to complete, before moving on.

    2. Let pendingTasks be this's global scopes's size.

    3. Let addedSuccessfully be false.

    4. For each workletGlobalScope of this's global scopes, queue a global task on the networking task source given workletGlobalScope to fetch a worklet script graph given moduleURLRecord, outsideSettings, this's worklet destination type, options["credentials"], workletGlobalScope's relevant settings object, this's module responses map, and the following steps given script:

      Only the first of these fetches will actually perform a network request; the ones for other WorkletGlobalScopes will reuse responses from this's module responses map.

      1. If script is null, then:

        1. Queue a global task on the networking task source given this's relevant global object to perform the following steps:

          1. If pendingTasks is not −1, then:

            1. Set pendingTasks to −1.

            2. Reject promise with an "AbortError" DOMException.

        2. Abort these steps.

      2. If script's error to rethrow is not null, then:

        1. Queue a global task on the networking task source given this's relevant global object to perform the following steps:

          1. If pendingTasks is not −1, then:

            1. Set pendingTasks to −1.

            2. Reject promise with script's error to rethrow.

        2. Abort these steps.

      3. If addedSuccessfully is false, then:

        1. Append moduleURLRecord to this's added modules list.

        2. Set addedSuccessfully to true.

      4. Run a module script given script.

      5. Queue a global task on the networking task source given this's relevant global object to perform the following steps:

        1. If pendingTasks is not −1, then:

          1. Set pendingTasks to pendingTasks − 1.

          2. If pendingTasks is 0, then resolve promise.

  6. Return promise.

11.3.3 The worklet's lifetime

The lifetime of a Worklet has no special considerations; it is tied to the object it belongs to, such as the Window.

Each Document has a worklet global scopes, which is a set of WorkletGlobalScopes, initially empty.

The lifetime of a WorkletGlobalScope is, at a minimum, tied to the Document whose worklet global scopes contain it. In particular, destroying the Document will terminate the corresponding WorkletGlobalScope and allow it to be garbage-collected.

Additionally, user agents may, at any time, terminate a given WorkletGlobalScope, unless the specification defining the corresponding worklet type says otherwise. For example, they might terminate them if the worklet agent's event loop has no tasks queued, or if the user agent has no pending operations planning to make use of the worklet, or if the user agent detects abnormal operations such as infinite loops or callbacks exceeding imposed time limits.

Finally, specifications for specific worklet types can give more specific details on when to create WorkletGlobalScopes for a given worklet type. For example, they might create them during specific processes that call upon worklet code, as in the example.