Skip to main content
Version: Next


First of all, thank you for your interest in Puppeteer! We'd love to accept your patches and contributions!

Contributor License Agreement

Contributions to this project must be accompanied by a Contributor License Agreement. You (or your employer) retain the copyright to your contribution, this simply gives us permission to use and redistribute your contributions as part of the project. Head over to <> to see your current agreements on file or to sign a new one.

You generally only need to submit a CLA once, so if you've already submitted one (even if it was for a different project), you probably don't need to do it again.

Getting started

  1. Clone this repository

    git clone
    cd puppeteer


    Open in GitHub Codespaces

  2. Install the dependencies

    npm install
    # Or to download Firefox by default
    PUPPETEER_PRODUCT=firefox npm install
  3. Build all packages

    npm run build
  4. Run all tests

    npm test

Building a single package

To build a single package, you can run:

npm run build --workspace <package> # e.g. puppeteer

This will build all dependent packages automatically, so specifying a single package is sufficient. This is all possible due to wireit which behaves similar to GNU Make.

Watch mode

To continuously build a package, you can run:

npm run build --watch --workspace <package> # e.g. puppeteer

You have to only specify a single package to watch else things will not work as expected As stated above because of wireit when a change happens all dependencies will be build or rebuild (if needed).

Removing stale artifacts

It's possible some generated artifacts (such as packages/puppeteer-core/src/types.ts) can become stale since these artifacts rely on complex conditions (such as names of distinct files) that cannot be captured by the build system. To clean artifacts, you can run

npm run clean
# or specify the package
npm run clean --workspace <package>

Comprehensive testing

Outside of npm test, there are several other npm scripts that are usually check through CI:

  • test-install - Tests whether puppeteer and puppeteer-core install properly and are functional.
  • test-types - Tests the TypeScript types in puppeteer using tsd.
  • test:chrome:** - Tests puppeteer on Chromium.
  • test:firefox:** - Tests puppeteer on Firefox.
  • unit - Runs unit tests.

The default npm test runs test:{chrome,firefox}:headless which is generally sufficient.

Puppeteer uses a custom test runner on top of Mocha that consults the TestExpectations.json to see if a given test result is expected or not. See more info about the test runner in tools/mocha-runner.

Unit tests

Tests that only test code (without the running browser) are put next to the classes they test and run using the Node test runner (requires Node 20+):

npm run unit

Code reviews

All submissions, including submissions by project members, require review. We use GitHub pull requests for this purpose. Consult GitHub Help for more information on using pull requests.

Code Style

Our coding style is fully defined in .eslintrc (ESLint) and .prettierrc.cjs (Prettier).

Code is checked for PRs automatically and you can check your code manually by running:

npm run lint

If some errors are returned, you can attempt to fix them using:

npm run format

Project structure

The following is a description of the primary folders in Puppeteer:

  • packages contains all public source code.
  • test contains all test source code.
  • test-d contains type tests using tsd.
  • tools contains miscellaneous scripts that are used in building and etc.
  • tools/mocha-runner - contains the source code for our test runner.

API guidelines

When authoring new API methods, consider the following:

  • Expose as little information as needed. When in doubt, don’t expose new information.
  • Methods are used in favor of getters/setters.
    • The only exception is namespaces, e.g. page.keyboard and page.coverage
  • All string literals must be small case. This includes event names and option values.
  • Avoid adding "sugar" API (API that is trivially implementable in user-space) unless they're extremely demanded.

Commit messages

Commit messages should follow the Conventional Commits format.

In particular, breaking changes should clearly be noted as “BREAKING CHANGE:” in the commit message footer. Example:

fix(page): fix method

This patch fixes so that it works with iframes.

Issues: #123, #234

BREAKING CHANGE: now delivers pizza at home by default.
To deliver to a different location, use the "deliver" option:
`{deliver: 'work'})`.

Writing documentation

Documentation is generated from TSDoc comments via npm run docs. It is automatically published to our documentation site on merge and gets versioned on release.

This means that you should not change the markdown in files docs/api manually.

Writing TSDoc comments

Each change to Puppeteer should be thoroughly documented using TSDoc comments. Refer to the API Extractor documentation for information on the exact syntax.

  • Every new method needs to have either @public or @internal added as a tag depending on if it is part of the public API.
  • Keep each line in a comment to no more than 90 characters (ESLint will warn you if you go over this). If you're a VSCode user the Rewrap plugin is highly recommended!

Running the documentation site locally

  1. At root, install all dependencies with npm i --ignore-scripts.
  2. run npm run docs which will generate all the .md files on puppeteer/docs/api.
  3. run npm i in puppeteer/website.
  4. run npm start in puppeteer/website.

Adding new dependencies

For all dependencies (both installation and development):

  • Do not add a dependency if the desired functionality is easily implementable.
  • If adding a dependency, it should be well-maintained and trustworthy.

A barrier for introducing new installation dependencies is especially high:

  • Do not add installation dependency unless it's critical to project success.

There are additional considerations for dependencies that are environment agonistic. See the third_party/ for details.

Testing tips

  • Every feature should be accompanied by a test.
  • Every public api event/method should be accompanied by a test.
  • Tests should not depend on external services.
  • Tests should work on all three platforms: Mac, Linux and Win. This is especially important for screenshot tests.

If a test is expected to fail on certain configurations or became flaky, update TestExpectations.json to reflect that. See more info about TestExpectations.json in tools/mocha-runner.

API Coverage

Every public API method or event should be called at least once in tests. To ensure this, the main test command runs coverage during testing.

Debugging Puppeteer

See Debugging Tips.

Debugging Puppeteer tests via VSCode

Copy the provided default .vscode/launch.template.json to .vscode/launch.json and then use the integrated VSCode debugger to debug test.

Remember to build test before launching via:

npm run build --workspace @puppeteer-test/test

For Project Maintainers

Rolling new Chrome version

There is a GitHub action that runs once per day. The action has a manual trigger that can be found on the Actions Tab.

Manual instructions

You can run the tools/update_browser_revision.mjs locally and try see if any changes need to be committed.

Note: You may need to run node --experimental-fetch tools/update_browser_revision.mjs as the script relies on fetch

The following steps are manual version of the script above.

  1. Find a suitable Chrome revision and version via or
  2. Update packages/puppeteer-core/src/revisions.ts with the found version number.
  3. Update versions.json with the new Chrome-to-Puppeteer version mapping and update lastMaintainedChromeVersion with the next one in from the list.
  4. Run npm run check. If it fails, update packages/puppeteer-core/package.json with the expected devtools-protocol version and run npm install to generate an updated package-lock.json.
  5. Run npm run clean, npm install and npm run build.
  6. Run npm test and ensure that all tests pass. If a test fails, bisect the upstream cause of the failure, and either update the test expectations accordingly (if it was an intended change) or work around the changes in Puppeteer (if it’s not desirable to change Puppeteer’s observable behavior).
  7. Commit and push your changes and open a pull request. The commit message must contain the version in Chrome <version> format to ensure that can parse it correctly, e.g. feat(chrome): roll to Chrome 90.0.4427.0.

Bisecting upstream changes

For bisecting Chrome/Chromium changes use

Releasing to npm

We use release-please to automate releases. When a release should be done, check for the release PR in our pull requests and merge it.

In case Release Please fails

In the event release-please fails, the following needs to be done:

  1. Update anything missing in the CHANGELOG of every package that was supposed to be published. For example, if the header is missing, you may need to add

    • For puppeteer:

    • For other packages:

  2. Create a GitHub release for each package, following the practice of previous releases.

Bug triage guidelines

Check incoming bug reports that do not have a confirmed or needs-feedback label:

  1. Make sure the issue is labeled as either bug or feature.
  2. If the issue does not have a clear repro or you cannot repro, ask for the repro and set the needs-feedback label.
  3. Follow-up on the issues you previously asked for a feedback on (you should get a notification on GitHub when the user responds).
  4. If the user does not provide feedback, the issue will be closed by the stale bot eventually.
  5. If you are able to reproduce the issue, add the label confirmed.
  6. If the bug is on the Chromium side, create a corresponding issue, label the GitHub issue with the upstream label, and post a link to in the comments.
  7. If the issue is not related to either Puppeteer or Chromium, close the issue.
  8. If the issue is about missing/incorrect documentation, label it as documentation.

Issues with PDFs:

  1. If the issue reproduces using the regular print dialog and/or headful, file a against the Blink>Layout component.
  2. If the issue is specific to Headless mode, file an issue on against the Internals>Headless component.