generated: '2026-08-09'
method: derived
source: openapi/webscraping-ai-openapi.yml
description: >-
  Entity graph derived from the published OpenAPI. WebScraping.AI is a stateless transformation
  API — it stores no user-owned resources and exposes no identifiers to reference. The only
  persistent entity is the caller's own Account; everything else is a per-request projection of
  a third-party web page. There are consequently no id-reference fields and no
  has_one/has_many relationships between server-side objects.
shape: stateless-transformation
entities:
- name: Account
  schema: '#/components/schemas/Account'
  kind: singleton
  persistent: true
  identifier: none (implicit — resolved from the api_key)
  read_by: [account]
  fields:
  - {name: email, type: string, description: Account email}
  - {name: remaining_api_calls, type: integer, deprecated: true, description: Deprecated alias
      of remaining_monthly_credits}
  - {name: remaining_monthly_credits, type: integer, description: Remaining subscription credits}
  - {name: remaining_payg_credits, type: integer, description: Remaining pay-as-you-go credits}
  - {name: remaining_total_credits, type: integer, description: Total remaining credits}
  - {name: resets_at, type: integer, description: Unix timestamp of the monthly quota reset}
  - {name: remaining_concurrency, type: integer, description: Concurrent-request headroom}
- name: SelectedAreas
  schema: '#/components/schemas/SelectedAreas'
  kind: response-projection
  persistent: false
  description: Array of HTML strings, one per element matched by the supplied selectors.
  produced_by: [getSelectedMultiple]
- name: Error
  schema: '#/components/schemas/Error'
  kind: envelope
  persistent: false
  description: Flat error envelope; carries the TARGET page's status alongside the message.
  produced_by: [getQuestion, getFields, getHTML, getText, getSelected, getSelectedMultiple, account]
  detail: errors/webscraping-ai-problem-types.yml
- name: TargetPage
  schema: null
  kind: external-subject
  persistent: false
  description: >-
    The third-party web page addressed by the `url` parameter. Not modeled in the spec and not
    owned by the provider, but it is the true subject of every operation — every response is a
    rendering of it (HTML, text, selected fragments, an AI answer, or extracted fields).
  addressed_by: [getQuestion, getFields, getHTML, getText, getSelected, getSelectedMultiple]
relationships:
- from: Account
  to: TargetPage
  type: none
  note: >-
    The API records no association between an account and the pages it scraped — no job,
    request or result resource is exposed for later retrieval.
- from: TargetPage
  to: SelectedAreas
  type: projection
  via: selectors
  note: Not a stored relationship; a per-request transformation.
representations:
- {operation: getHTML, of: TargetPage, media_type: text/html, note: fully rendered markup}
- {operation: getText, of: TargetPage, media_type: 'text/html | application/json', note: 'visible
    text; json/xml add title, description, content'}
- {operation: getSelected, of: TargetPage, media_type: text/html, note: one selected area}
- {operation: getSelectedMultiple, of: TargetPage, media_type: application/json, schema: SelectedAreas}
- {operation: getQuestion, of: TargetPage, media_type: text/html, note: LLM answer as plain text}
- {operation: getFields, of: TargetPage, media_type: application/json, note: 'caller-defined field
    map; the shape is supplied per request, so it cannot be modeled as a fixed schema'}
- {operation: account, of: Account, media_type: application/json, schema: Account}
observations:
- No id-prefix convention exists, because no server-side object is ever assigned an id.
- The response shape of getFields is defined by the caller's `fields` argument, which is why the
  spec cannot declare a concrete schema for it — a real constraint for agents doing typed
  deserialization.
- Credit accounting is the only stateful concept, and it is readable through one operation.
x-evidence:
  fetched: '2026-08-09'
  probes:
  - {url: 'https://webscraping.ai/openapi.yml', http_status: 200}