Skip to content
Workers
Visit Workers on GitHub
Set theme to dark (⇧+D)

Configuration

Background

Your project will need some configuration before you can publish your Worker. Configuration is done through changes to keys and values stored in a wrangler.toml file located in the root of your project directory. You must manually edit this file to edit your keys and values before you can publish.


Environments

The top-level configuration is the collection of values you specify at the top of your wrangler.toml file. These values will be inherited by all environments, unless otherwise defined in the environment.

The layout of a top-level configuration in a wrangler.toml file is displayed below:

name = "your-worker"
type = "javascript"
account_id = "your-account-id"
# This field specifies that the Worker
# will be deployed to a *.workers.dev domain
workers_dev = true
# -- OR --
# These fields specify that the Worker
# will deploy to a custom domain
zone_id = "your-zone-id"
routes = ["example.com/*"]

Environment configuration (optional): the configuration values you specify under an [env.name] in your wrangler.toml file.

Environments allow you to deploy the same project to multiple places under multiple names. These environments are utilized with the --env or -e flag on the commands that are deploying live Worker scripts:

  • build
  • dev
  • preview
  • publish
  • secret

Some environment properties can be inherited from the top-level configuration, but if new values are configured in an environment, they will always override those at the top level.

An example of an [env.name] configuration looks like this:

name = "your-worker"
type = "javascript"
account_id = "your-account-id"
[env.helloworld]
# These new values will override the top level configuration.
name = "your-worker-helloworld"
account_id = "your-other-account-id"
# Any additional keys, like environment variables, will be placed here.
vars = { FOO = "some value", BAR = "some other string" }
kv_namespaces = [
{ binding = "FOO", id = "1a2b3c4d5e", preview_id = "6e7f8g9h10i" }
]

To deploy this example Worker to the helloworld environment, you would run wrangler publish --env helloworld.


Keys

There are three types of keys in a wrangler.toml file:

  • Top level only keys are required to be configured at the top level of your wrangler.toml file only; multiple environments on the same project must share this key's value.

  • Inherited keys can be configured at the top level and/or environment. If the key is defined only at the top level, the environment will use the key's value from the top level. If the key is defined in the environment, the environment value will override the top-level value.

  • Non-inherited keys must be defined for every environment individually.

  • name inherited required

    • The name of your Worker script. If inherited, your environment name will be appended to the top level.
  • type top level required

    • Specifies how wrangler build will build your project. There are three options: javascript, webpack, and rust. javascript checks for a build command specified in the [build] section, webpack builds your project using webpack v4, and rust compiles the Rust in your project to WebAssembly.
  • account_id inherited required

    • This is the ID of the account associated with your zone. You might have more than one account, so make sure to use the ID of the account associated with the zone_id you provide, if you provide one. It can also be specified through the CF_ACCOUNT_ID environment variable.
  • zone_id inherited optional

    • This is the ID of the zone or domain you want to run your script on. It can also be specified through the CF_ZONE_ID environment variable. This key is optional if you are using only a *.workers.dev subdomain.
  • workers_dev inherited optional

  • route not inherited optional

    • A route, specified by URL pattern, on your zone that you would like to run your Worker on.
      route = "http://example.com/*". A route OR routes key is only required if you are not using a *.workers.dev subdomain.
  • routes not inherited optional

    • A list of routes you would like to use your Worker on. These follow exactly the same rules a route, but you can specify a list of them.
      routes = ["http://example.com/hello", "http://example.com/goodbye"]. A route OR routes key is only required if you are not using a *.workers.dev subdomain.
  • webpack_config inherited optional

    • This is the path to a custom webpack configuration file for your Worker. You must specify this field to use a custom webpack configuration, otherwise Wrangler will use a default configuration for you. Refer to the Wrangler webpack page for more information.
  • vars not inherited optional

    • An object containing text variables that can be directly accessed in a Worker script.
  • kv_namespaces not inherited optional

    • These specify any Workers KV Namespaces you want to access from inside your Worker.
  • site inherited optional

    • Determines the local folder to upload and serve from a Worker.
  • dev not inherited optional

    • Arguments for wrangler dev that configure local server.
  • triggers inherited optional

    • Configures cron triggers for running a Worker on a schedule.
  • usage_model inherited optional

  • build top level optional

    • Configures a custom build step to be run by Wrangler when building your Worker. Refer to the custom builds documentation for more details.

vars

The vars key defines a table of environment variables provided to your Worker script. All values are plaintext values.

Usage:

[vars]
FOO = "some value"
BAR = "some other string"

The table keys are available to your script as global variables, which will contain their associated values.

// Worker code:
console.log(FOO);
//=> "some value"
console.log(BAR);
//=> "some other string"

Alternatively, you can define vars using an inline table format. This style should not include any new lines to be considered a valid TOML configuration:

vars = { FOO = "some value", BAR = "some other string" }

kv_namespaces

kv_namespaces defines a list of KV namespace bindings for your Worker.

Usage:

kv_namespaces = [
{ binding = "FOO", id = "0f2ac74b498b48028cb68387c421e279", preview_id = "6a1ddb03f3ec250963f0a1e46820076f" },
{ binding = "BAR", id = "068c101e168d03c65bddf4ba75150fb0", preview_id = "fb69528dbc7336525313f2e8c3b17db0" }
]

Alternatively, you can define kv namespaces like so:

[[kv_namespaces]]
binding = "FOO"
preview_id = "abc456"
id = "abc123"
[[kv_namespaces]]
binding = "BAR"
preview_id = "xyz456"
id = "xyz123"

Much like environment variables and secrets, the binding names are available to your Worker as global variables.

// Worker script:
let value = await FOO.get("keyname");
//=> gets the value for "keyname" from
//=> the FOO variable, which points to
//=> the "0f2ac...e279" KV namespace
  • binding required

    • The name of the global variable your code will reference. It will be provided as a KV runtime instance.
  • id required

    • The ID of the KV namespace that your binding should represent. Required for wrangler publish.
  • preview_id required

    • The ID of the KV namespace that your binding should represent during wrangler dev or wrangler preview. Required for wrangler dev and wrangler preview.

site

A Workers Site generated with wrangler generate --site or wrangler init --site.

Usage:

[site]
bucket = "./public"
entry-point = "workers-site"
  • bucket required

    • The directory containing your static assets. It must be a path relative to your wrangler.toml file. Example: bucket = "./public"
  • entry-point optional

    • The location of your Worker script. The default location is workers-site. Example: entry-point = "./workers-site"
  • include optional

    • An exclusive list of .gitignore-style patterns that match file or directory names from your bucket location. Only matched items will be uploaded. Example: include = ["upload_dir"]
  • exclude optional

    • A list of .gitignore-style patterns that match files or directories in your bucket that should be excluded from uploads. Example: exclude = ["ignore_dir"]

You can also define your site using an alternative TOML syntax.

Storage Limits

For exceptionally large pages, Workers Sites may not be ideal. There is a 25MB limit per page or file. Additionally, Wrangler will create an asset manifest for your files that will count towards your script’s size limit. If you have too many files, you may not be able to use Workers Sites.

Exclusively including files/directories

If you want to include only a certain set of files or directories in your bucket, add an include field to your [site] section of your wrangler.toml file:

[site]
bucket = "./public"
entry-point = "workers-site"
include = ["included_dir"] # must be an array.

Wrangler will only upload files or directories matching the patterns in the include array.

Excluding files/directories

If you want to exclude files or directories in your bucket, add an exclude field to your [site] section of your wrangler.toml file:

[site]
bucket = "./public"
entry-point = "workers-site"
exclude = ["excluded_dir"] # must be an array.

Wrangler will ignore files or directories matching the patterns in the exclude array when uploading assets to Workers KV.

Include > Exclude

If you provide both include and exclude fields, the include field will be used and the exclude field will be ignored.

Default ignored entries

Wrangler will always ignore:

  • node_modules
  • Hidden files and directories
  • Symlinks

More about include/exclude patterns

Refer to the gitignore documentation to learn more about the standard matching patterns.

Customizing your Sites Build

Workers Sites projects use webpack by default. Though you can bring your own webpack configuration, be aware of your entry and context settings.

You can also use the [build] section with Workers Sites, as long as your build step will resolve dependencies in node_modules. Refer to the custom builds section for more information.

triggers

A set of cron triggers used to call a Worker on a schedule.

Usage:

[triggers]
crons = ["0 0 * JAN-JUN FRI", "0 0 LW JUL-DEC *"]

dev

Arguments for wrangler dev can be configured here so you do not have to repeatedly pass them.

Usage:

[dev]
port = 9000
local_protocol = "https"
  • ip optional

    • IP address for the local wrangler dev server to listen on, defaults to 127.0.0.1.
  • port optional

    • Port for local wrangler dev server to listen on, defaults to 8787.
  • local_protocol optional

    • Protocol that local wrangler dev server listen to requests on, defaults to http.
  • upstream_protocol optional

    • Protocol that wrangler dev forwards requests on, defaults to https.

build

A custom build command for your project. There are two configurations based on the format of your Worker: service-worker and modules (beta).

Service Workers

This section is for customizing Workers with the service-worker format. These Workers use addEventListener and look like the following:

addEventListener('fetch', event => {
event.respondWith(new Response("I'm a service Worker!"))
})

Usage:

[build]
command = "npm install && npm run build"
[build.upload]
format = "service-worker"
[build]
  • command optional

    • The command used to build your Worker. On Linux and macOS, the command is executed in the sh shell and the cmd shell for Windows. The && and || shell operators may be used.
  • cwd optional

    • The working directory for commands, defaults to the project root directory.
  • watch_dir optional

    • The directory to watch for changes while using wrangler dev, defaults to the src relative to the project root directory.
[build.upload]
  • format required

    • The format of the Worker script, must be "service-worker".

Modules

Workers now supports the ES Modules syntax. Modules support in Cloudflare Workers is currently in beta. This format allows you to export a collection of files and/or modules, unlike the Service Worker format which requires a single file to be uploaded.

Module Workers export their event handlers instead of using addEventListener calls.

Modules receive all bindings (KV Namespaces, Environment Variables, and Secrets) as arguments to the exported handlers. With the Service Worker format, these bindings are available as global variables.

An uploaded module may import other uploaded ES Modules. If using the CommonJS format, you may require other uploaded CommonJS modules.

import html from './index.html'
export default {
// * request is the same as `event.request` from the service worker format
// * waitUntil() and passThroughOnException() are accessible from `ctx` instead of `event` from the service worker format
// * env is where bindings like KV namespaces, Durable Object namespaces, Config variables, and Secrets
// are exposed, instead of them being placed in global scope.
async fetch(request, env, ctx) {
const headers = { 'Content-Type': 'text/html;charset=UTF-8' }
return new Response(html, { headers })
}
}

To create a Workers project using Wrangler and Modules, add a [build] section:

[build]
command = "npm install && npm run build"
[build.upload]
format = "modules"
main = "./worker.mjs"
[build]
  • command optional

    • The command used to build your Worker. On Linux and macOS system, the command is executed in the sh shell and the cmd shell for Windows. The && and || shell operators may be used.
  • cwd optional

    • The working directory for commands, defaults to the project root directory.
  • watch_dir optional

    • The directory to watch for changes while using wrangler dev, defaults to the src relative to the project root directory.
[build.upload]
  • format required

    • The format of the Workers script, must be "modules".
  • dir optional

    • The directory you wish to upload your modules from, defaults to the dist relative to the project root directory.
  • main required

    • The relative path of the main module from dir, including the ./ prefix. The main module must be an ES module. For projects with a build script, this usually refers to the output of your JavaScript bundler.
  • rules optional

    • An ordered list of rules that define which modules to import, and what type to import them as. You will need to specify rules to use Text, Data, and CompiledWasm modules, or when you wish to have a .js file be treated as an ESModule instead of CommonJS.

    • Defaults:

      [build.upload]
      format = "modules"
      main = "./worker.mjs"
      # You do not need to include these default rules in your `wrangler.toml`, they are implicit.
      # The default rules are treated as the last two rules in the list.
      [[build.upload.rules]]
      type = "ESModule"
      globs = ["**/*.mjs"]
      [[build.upload.rules]]
      type = "CommonJS"
      globs = ["**/*.js", "**/*.cjs"]
      • type required

        • The module type, see the table below for acceptable options
        typeJavaScript type
        ESModule-
        CommonJS-
        TextString
        DataArrayBuffer
        CompiledWasmWebAssembly.Module
      • globs required

        • Unix-style glob rules that are used to determine the module type to use for a given file in dir. Globs are matched against the module's relative path from build.upload.dir without the ./ prefix. Rules are evaluated in order, starting at the top.
      • fallthrough optional

        • This option allows further rules for this module type to be considered if set to true. If not specified or set to false, further rules for this module type will be ignored.

Example

To illustrate how these levels are applied, here is a wrangler.toml file using multiple environments:

wrangler.toml
# top level configuration
type = "javascript"
name = "my-worker-dev"
account_id = "12345678901234567890"
zone_id = "09876543210987654321"
route = "dev.example.com/*"
usage_model = "unbound"
kv_namespaces = [
{ binding = "FOO", id = "b941aabb520e61dcaaeaa64b4d8f8358", preview_id = "03c8c8dd3b032b0528f6547d0e1a83f3" },
{ binding = "BAR", id = "90e6f6abd5b4f981c748c532844461ae", preview_id = "e5011a026c5032c09af62c55ecc3f438" }
]
[build]
command = "webpack"
[build.upload]
format = "service-worker"
[site]
bucket = "./public"
entry-point = "workers-site"
[dev]
ip = "0.0.0.0"
port = 9000
local_protocol="http"
upstream_protocol="https"
# environment configuration
[env.staging]
name = "my-worker-staging"
route = "staging.example.com/*"
kv_namespaces = [
{ binding = "FOO", id = "0f2ac74b498b48028cb68387c421e279" },
{ binding = "BAR", id = "068c101e168d03c65bddf4ba75150fb0" }
]
# environment configuration
[env.production]
workers_dev= true
kv_namespaces = [
{ binding = "FOO", id = "0d2ac74b498b48028cb68387c421e233" },
{ binding = "BAR", id = "0d8c101e168d03c65bddf4ba75150f33" }
]