generated: '2026-08-14'
method: derived
source: openapi/contextdev-openapi.yml, openapi/contextdev-batch-api-openapi.yml
note: 'Context.dev is mostly a stateless enrichment/extraction API (Brand, Web, Person, Classification
  responses are computed, not stored). Two persistent, identified resource graphs exist: the Monitors
  subsystem, and — added 2026-08-14 — the Batch subsystem. They are entirely disjoint: no field links
  a Batch to a Monitor or vice versa, and they are even rate-limited from separate buckets.'
entities:
- name: Monitor
  schema: MonitorsMonitor
  id: id
  description: A configured watch over a page, sitemap, or extracted dataset.
- name: MonitorRun
  schema: MonitorsRun
  id: id
  description: One execution of a monitor; may or may not produce a change.
- name: MonitorChange
  schema: MonitorsChange
  id: id
  description: A detected change (diff, evidence, confidence, importance).
- name: Brand
  schema: BrandData
  id: domain
  stateless: true
  description: Computed brand profile (logos, colors, socials, industry) for a domain.
- name: Person
  schema: PeopleData
  stateless: true
  description: Computed person/contact enrichment.
- name: Product
  schema: ProductData
  stateless: true
  description: Structured product record extracted from a URL.
- name: Batch
  schema: Batch
  id: id
  id_prefix: batch_
  description: An asynchronous scrape or crawl job over up to 25,000 URLs. Carries status (queued, running,
    cancelling, completed, cancelled, failed — the last three terminal), mode (scrape | crawl), format
    (markdown | html), and tags stored at submission.
  added: '2026-08-14'
- name: BatchDetail
  schema: BatchDetail
  id: id
  description: The expanded batch record returned by getBatch, including progress counters, per-page error
    counts and download links once the batch settles.
  added: '2026-08-14'
- name: BatchResultRecord
  schema: BatchResultRecord
  description: One scraped page inside a finished batch — content plus PageMetadata. Retrieved page-by-page
    via getBatchResults, or in bulk as gzipped NDJSON.
  added: '2026-08-14'
- name: PageMetadata
  schema: PageMetadata
  stateless: true
  description: Per-page metadata (headings, alternates, values) attached to a batch result record.
  added: '2026-08-14'
relationships:
- from: MonitorRun
  to: Monitor
  type: belongs_to
  via: monitor_id
- from: Monitor
  to: MonitorRun
  type: has_many
  via: monitor_id
- from: MonitorChange
  to: Monitor
  type: belongs_to
  via: monitor_id
- from: MonitorChange
  to: MonitorRun
  type: belongs_to
  via: run_id
- from: MonitorRun
  to: MonitorChange
  type: has_one
  via: change_id
- from: BatchResultRecord
  to: Batch
  type: belongs_to
  via: batch_id
  added: '2026-08-14'
- from: Batch
  to: BatchResultRecord
  type: has_many
  via: batch_id
  added: '2026-08-14'
- from: BatchDetail
  to: Batch
  type: has_one
  via: id
  note: Same resource; BatchDetail is the expanded read projection.
  added: '2026-08-14'
- from: BatchResultRecord
  to: PageMetadata
  type: has_one
  via: metadata
  added: '2026-08-14'
id_prefixes:
  Batch: batch_ (example batch_9f2c8a)
  note: Only the Batch entity documents a typed id prefix in the spec; Monitor, MonitorRun and MonitorChange
    ids carry no documented prefix.
subsystems:
- name: Monitors
  entities:
  - Monitor
  - MonitorRun
  - MonitorChange
  persistent: true
  rate_limit_bucket: per-organization, 1000 req/min, flat across plans
- name: Batch
  entities:
  - Batch
  - BatchDetail
  - BatchResultRecord
  - PageMetadata
  persistent: true
  rate_limit_bucket: per-plan per-minute cap; submission costs 1 request regardless of size
  retention: Results are stored until deleted via deleteBatch; no automatic retention window is documented.
- name: Enrichment / extraction
  entities:
  - Brand
  - Person
  - Product
  persistent: false
  note: Computed per call, not stored, not addressable by id.