Skip to main content
Orient yourself in 30 seconds. Each row is one endpoint; click through for the full parameter reference. This page covers the endpoints the recipes use plus the main endpoint groups; the complete list (including Documents, Knowledge bases, Projects, Integrations, and OpenAI Ads) is in the REST API tab. For deeper context on conventions (dates, ordering, pagination), see Conventions & gotchas.

​
Organization & setup

​
Reports

Answer Engine Insights report endpoints are POST requests scoped by category_id and return the {info, data} shape. Visibility, citations, sentiment, and query-fanouts reports take category_id, start_date, end_date, and metrics, plus report-specific optional fields. Traffic reports use a different shape: bots and referrals are scoped by domain rather than category_id. V2 FactCheck reports and V2 prompt answers are category-and-date scoped, but do not take a metrics field.

​
/v1/reports/visibility

How often your asset appears in AI answers. Common recipes: Visibility Score (with delta), over time, leaderboard, compare competitors, segment by model/region/persona.

​
/v1/reports/citations

Which URLs and domains AI answers cite. Common recipes: Citation Share + delta, share + volume over time, rank by domain.
citation_share is per-AI-model averaged. count is the raw number. Sorting by one doesn’t track the other. Pick the question you’re answering and sort on the matching metric.

​
/v1/reports/sentiment

How positively/negatively AI answers reference your asset.
/v1/reports/sentiment is the legacy sentiment processor and its numbers will not match the Profound app’s headline value. The app reads from Sentiment v2: use POST /v2/reports/sentiment (see Filtering & concepts). Its positive_sentiment / negative_sentiment metrics are percentages that sum to 100 per row, so they are not comparable with v1’s counts.

​
/v1/reports/query-fanouts

The query-fanout view (how prompts cascade across AI surfaces). Same {info, data} shape.

​
/v1/reports/referrals (and /v2/reports/referrals)

Human-referral traffic to your domains attributed back to AI answers.

​
/v1/reports/bots (and /v2/reports/bots)

AI crawler / bot traffic to your domains.

​
/v2/reports/factcheck and /v2/reports/factcheck/claims

FactCheck scores and inaccurate claims for a category. These endpoints take category_id, start_date, and end_date, but not metrics.

​
V2 report endpoints

POST /v2/reports/visibility, /v2/reports/citations, /v2/reports/sentiment, and /v2/reports/query-fanouts are the V2 equivalents of the V1 reports above, with an inclusive end_date, a filter tree, named metric fields on each row, and cursor pagination. /v2/reports/shopping/* and /v2/reports/social/youtube/* are V2-only. Each has a /stream SSE variant. See Filtering & concepts.

​
Answers (per-prompt data)

​
Agents

​
Prompts

​
Content Optimization (beta)

For the canonical parameter reference (every operator, every field, required vs optional), see the REST API tab.