Browserless Full Page Archive

arazzo: 1.0.1
info:
  title: Browserless Full Page Archive
  summary: Unblock a protected URL, then branch into structured scraping or a content+PDF archive depending on whether unblocked HTML was returned.
  description: >-
    A branching archival pipeline. The workflow first calls /chrome/unblock to
    clear bot detection on the target URL, returning cookies, HTML content, and
    a base64 screenshot. It then branches on whether the unblock step returned
    usable HTML content: when content is present it runs /chrome/scrape with the
    recovered cookies to extract structured element data, and when content is
    absent it falls back to a /chrome/content render so the archive always
    captures the page HTML. Both branches reuse the cleared cookies. Every step
    spells out its token query parameter and JSON request body inline so the
    flow can be read and executed without opening the underlying OpenAPI
    description.
  version: 1.0.0
sourceDescriptions:
- name: browserlessApi
  url: ../openapi/browserless-openapi.yml
  type: openapi
workflows:
- workflowId: full-page-archive
  summary: Unblock a URL, then scrape structured data or fall back to a content render.
  description: >-
    Clears bot detection via /chrome/unblock, then branches: when unblocked HTML
    content is present it scrapes structured elements via /chrome/scrape using
    the recovered cookies, otherwise it falls back to a /chrome/content render
    using the same cookies. Both branches reuse the cleared session.
  inputs:
    type: object
    required:
    - token
    - url
    - selectors
    properties:
      token:
        type: string
        description: The Browserless authorization token passed as a query parameter.
      url:
        type: string
        description: The URL to unblock and archive.
      selectors:
        type: array
        description: An array of element selector objects to extract when content is available.
        items:
          type: object
          required:
          - selector
          properties:
            selector:
              type: string
              description: A CSS selector identifying the elements to extract.
            timeout:
              type: number
              description: Maximum time in milliseconds to wait for the selector.
  steps:
  - stepId: unblockSite
    description: >-
      Clear bot detection on the URL, returning cookies, HTML content, and a
      base64 full-page screenshot.
    operationPath: '{$sourceDescriptions.browserlessApi.url}#/paths/~1chrome~1unblock/post'
    parameters:
    - name: token
      in: query
      value: $inputs.token
    requestBody:
      contentType: application/json
      payload:
        url: $inputs.url
        cookies: true
        content: true
        screenshot: true
        bestAttempt: true
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      cookies: $response.body#/cookies
      content: $response.body#/content
      screenshot: $response.body#/screenshot
    onSuccess:
    - name: hasContent
      type: goto
      stepId: scrapeElements
      criteria:
      - context: $response.body
        condition: $.content != null
        type: jsonpath
    - name: noContent
      type: goto
      stepId: fallbackContent
      criteria:
      - context: $response.body
        condition: $.content == null
        type: jsonpath
  - stepId: scrapeElements
    description: >-
      Extract structured element data for the supplied selectors from the
      unblocked URL, reusing the cookies recovered from the unblock step.
    operationPath: '{$sourceDescriptions.browserlessApi.url}#/paths/~1chrome~1scrape/post'
    parameters:
    - name: token
      in: query
      value: $inputs.token
    requestBody:
      contentType: application/json
      payload:
        url: $inputs.url
        elements: $inputs.selectors
        cookies: $steps.unblockSite.outputs.cookies
        bestAttempt: true
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      data: $response.body#/data
    onSuccess:
    - name: done
      type: end
  - stepId: fallbackContent
    description: >-
      Fall back to rendering the page HTML via /chrome/content when the unblock
      step returned no content, reusing the recovered cookies.
    operationPath: '{$sourceDescriptions.browserlessApi.url}#/paths/~1chrome~1content/post'
    parameters:
    - name: token
      in: query
      value: $inputs.token
    requestBody:
      contentType: application/json
      payload:
        url: $inputs.url
        cookies: $steps.unblockSite.outputs.cookies
        bestAttempt: true
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      html: $response.body
  outputs:
    unblockScreenshot: $steps.unblockSite.outputs.screenshot
    scrapedData: $steps.scrapeElements.outputs.data
    fallbackHtml: $steps.fallbackContent.outputs.html