Skip to content

Open-source agentic AI Search Execution Platform. Track AI visibility, find opportunities, and let agents turn signals into prioritized actions and drafted fixes across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot.

1.4 KB
Raw Blame
-- ============================================================================
-- 00274: Quality phase 2, CE-050 — the page read's checks
-- ============================================================================
-- The page read (00270) also says, in the page's own words:
--
--   * answerFirst: whether the paragraph right after the H1 answers it;
--   * recommends: where the page puts the brand forward among options;
--   * disclosure: where it says the brand makes what it recommends;
--   * balance: a real limit of the brand's own product, or where another
--     option wins;
--   * how: how the page's content was made (who, how, why);
--   * cut: whether the model read less than the whole page (a check not
--     found may be in the rest).
--
-- Each found one names its paragraph by its hash, as the read's items do;
-- null when the page wasn't read for them (a read before this, a missed one).

alter table public.content_quality_pages
  add column if not exists checks jsonb
    check (
      checks is null
      or (jsonb_typeof(checks) = 'object' and octet_length(checks::text) <= 4096)
    );

comment on column public.content_quality_pages.checks is
  'Quality CE-050 (00274): { answerFirst: { block_hash, answers } | null, recommends | disclosure | balance | how: { quote, block_hash } | null, cut: boolean }; null when the page wasn''t read for them.';

notify pgrst, 'reload schema';