Skip to content
AdlyseDocs
Product guideMCP reference
⌘K
Open Adlyse
⌘K
Overview
QuickstartAuthentication
adlyse-insights
list_ad_accountsget_dashboard_snapshotget_impact_analysisget_competitor_pressurelist_guidelinesget_brand_kitupdate_brand_kitsave_brand_assetget_guidelinelist_suggestionsget_suggestionget_strategy_sessionlist_strategy_claimslist_strategy_proposalsget_creative_insightslist_watched_competitorslist_competitor_adswatch_competitorunwatch_competitorlist_audit_eventslist_organic_propertiesget_organic_overviewget_organic_page_metricsinspect_pageget_organic_performanceget_site_inventorycheck_site_healthget_ai_visibilitylist_ai_visibility_answerslist_ai_visibility_promptsupsert_ai_visibility_promptsrun_ai_visibility_checksget_growth_evidencelist_site_connection_optionslist_site_key_eventssave_site_sourceremove_site_sourceresync_site_sourceupdate_growth_deliveryget_growth_deliveryget_geo_workspaceupdate_geo_workspace
adlyse-ads
adlyse-workflows
adlyse-admin
adlyse-all
  1. Overview
  2. adlyse-insights

get_geo_workspace

Read the selected site’s GEO workspace. Requires an entitled organization and dashboard permission. site_source_id is required and must belong to that organization. research=true returns observed GSC/paid-search candidates with provenance instead of the workspace; search demand is a proxy, not AI volume.

The default result is compact: revision, settings, provider capabilities (each with last_run_error when the latest run could not ask that engine), organization budget, reviewed entities (the organization’s brands included in AI checks, shared by every site), approved facts (with facts_revision), claim assessments and their stale flag, existing work (the site’s cards, including ai_source_gap and ai_brand_claim_contradiction findings), competitor_candidates, asynchronous portfolio_review, crawl and a notice. omitted_sections names the rest; pass sections to read any of them by name: question portfolio, source domains, weekly history, competitive_results, perception, recent answers, eligible members, audit events, page outcomes, page_health, organization sites coverage and the organization’s whole brands list (each with in_ai_checks, watching_ads and source). limit (default 25, max 250) caps every list section and offset (echoed in the result) skips rows in every one of them, so page one section at a time by asking for it alone. truncated maps each paged or cut section (including crawl.rows) to its full row count. The result is also trimmed to fit one agent observation (longest list first), so a page may come back shorter than limit: the next page starts at offset plus the rows actually returned, not at offset + limit.

Evidence is bounded: 180 days/10,000 probes, 100 recent answers, 100 domain groups, 100 events, 200 active facts plus 50 retired facts, 200 claims, 50 sites, 50 health observations, 50 crawl.rows (its requests count covers up to 1,000), 20 cited landing pages. history_truncated identifies the probe limit. Exports share these bounds. Do not present this as an exhaustive census of AI answers or a site-wide causal revenue attribution.

Use the current revision when applying a reviewed change. Error responses contain error and error_type; unavailable analytics is not zero traffic.

Previousget_growth_deliveryNextupdate_geo_workspace