Skip to main content
Returns AI-citation attribution for posted content. Computed from nightly snapshots and normalized, so frequently-run prompts don’t drown out the signal. Three signals per post: lifetime totals, recent counts within the window (default 14d), and a weekly trajectory series for sparklines. Lifetime totals since posted: citing_runs = answers that cited the post; brand_answers = the brand’s answers across all its tracked questions. Also returns per-platform breakdowns and the top prompts driving citations. Use this when the user asks:
  • “What’s working — which posts drive AI citations?”
  • “Are my posts getting cited more over time?”
  • “Which prompts is this LinkedIn post showing up on?”
  • “How does this blog compare to that one for AI visibility?”
Skip when the user wants GA4 web traffic for posted pages (use get_content_performance) or AI visibility across the brand at large (use get_brand_visibility). Engines: each post has per-engine figures in platforms.lifetime and platforms.recent (keys: chatgpt, google_aio, perplexity, gemini, claude, grok; only engines the brand is scanned on). citing = answers that cited the post; brand_answers = every answer to every question the brand tracks on that engine, not only this post’s questions. So quote answer counts per engine (“cited in 12 Google AI Overviews answers since <platforms.lifetime[k].since>”), never a rate or a fraction: against that base a cited post looks like it is failing. Recent window: recent.window_days days ending at recent.window_end, the newest day the nightly job has counted (it never counts today, so this lags by a day or two). In platforms.recent, an engine with brand_answers 0 did not run in the window: citing is null and checked is false. Say it wasn’t checked, never that it cited the post 0 times. When recent.days_with_data is 0, recent.citing_runs is null: nothing was counted for this post in the window. Say it wasn’t checked, never that it wasn’t cited. recent.days_with_data is how many of those days have data for the post; when it is lower than window_days, say so (“cited in 3 answers in the 14 days to 2026-10-16, 9 of them with data”). Pass content_id to narrow to a single posted content row. Omit for the brand-wide top 200 most recent posts. The response includes posts with no snapshots yet (zero baseline) so the caller can render an empty state without a second round-trip.

​
Inputs