Pydoll adapter¶

Install the optional browser dependency:

python -m pip install -e ".[browser]"

Enable Pydoll in a workflow:

browser:
  engine: pydoll
  headless: true
  start_timeout: 20
  default_wait_timeout: 10
  network_idle_period: 0.5
  download_directory: output/downloads
  arguments:
    - --window-size=1440,900

binary_location may be an absolute path or a path relative to the workflow file. The adapter starts Chromium lazily, so a workflow with browser settings but no browser steps does not launch a process.

Operation mapping¶

Crawlerflow operation	Pydoll API
`goto`	`Tab.go_to()`
`click`, `fill`, `wait`	`Tab.query()` and `WebElement` methods
`evaluate`	`Tab.execute_script()`
`cookies`, `set_cookies`	`Tab.get_cookies()`, `Tab.set_cookies()`
`browser_request`	`Tab.request.request()`
`download`	Browser-context `GET` request
`screenshot`	`Tab.take_screenshot()`

wait_network subscribes to CDP request-started, loading-finished, and loading-failed events. It returns after no tracked request remains active for network_idle_period seconds.

Cookies and artifacts¶

steps:
  - get_cookies:
      save_as: session_cookies
  - set_cookies:
      cookies: "{{session_cookies}}"
  - download:
      url: https://example.com/report.csv
      path: output/report.csv
      save_as: downloaded_file
  - screenshot:
      path: output/page.png
      save_as: screenshot_file

Relative paths resolve from the workflow file's directory. get_cookies also updates the shared workflow cookie state; download and screenshot return their resolved path and optionally expose it through save_as.