CitationLab Marketing Manager

Gets Cite Analytics overview with daily stats for a project

get_analytics_overview

Gets GA4 conversion/event data showing event counts, users, and event values, totalled per event name. It has no dimension filter: to split ONE event by source, medium, page or an event parameter (e.g. form_location), use analytics.event_breakdown (source=ga4; dimension session_source, session_medium, page_path, device_category, country or properties). The source/medium/page/device/country splits are cached only for events that are an active GA4 conversion definition or a funnel step, so for any other event they come back empty (not zero data: make the event a conversion first); properties and event_property_values read GA4 live and analytics.event_property_values (the values of one registered custom dimension, with an optional date range).

get_ga4_events

Gets GA4 page-level engagement data showing pageviews, engagement rate, time on page, bounce rate, entrances per page

get_ga4_page_engagement

Gets a comprehensive GA4 analytics summary for a project including totals, top sources, top pages, device breakdown, and top countries

get_ga4_summary

Gets GA4 traffic sources breakdown by source/medium/channel with sessions, users, bounce rate, and conversions

get_ga4_traffic_sources

Gets Google Ads metrics by entity type (campaign, ad_group, ad, keyword, search_term, or auction_insight). There is no account/customer-level entity type — for account totals, aggregate the campaign rows.

get_gads_metrics

Gets KPI trend data (daily values) over a period

get_kpi_trend

Gets LinkedIn Ads performance metrics for a project

get_linkedin_ads_metrics

Get Meta (Facebook/Instagram) Ads daily performance rows for a project — the right source for week-over-week or day-level comparisons. Populated by meta_ads_sync / the 12h scheduler; empty until at least one sync has run.

get_meta_ads_daily_metrics

Get Meta Ads monthly performance metrics for a project. For day/week-level numbers (e.g. week-over-week comparisons) use get_meta_ads_daily_metrics instead — this tool only returns one row per entity per calendar month.

get_meta_ads_metrics

Get Bing Ads performance metrics for a project.

get_bing_ads_metrics

Gets Google Search Console page performance data. When `startDate`/`endDate` are given explicitly, the response is scoped to exactly that period (fetched live from Google — see `source`). With no explicit period, it serves the latest synced snapshot for speed, which is a rolling ~400-day cumulative window rather than a fixed period — check `note`/`asOf` to tell the two cases apart.

get_gsc_pages

Gets Google Search Console keyword/query data aggregated across all days in the period. Returns accurate totals plus top keywords by impressions. Check `dataCompleteness.isComplete` before trusting the totals as covering the whole requested period — `missingDates` lists any day inside the reporting window (GSC's normal ~3-day lag) that never synced, and `lastSyncedAt` says when this site's data was last touched by a sync. Pass 'keywords' to filter to specific exact queries and get a per-page breakdown (which URL ranks for each). Pass 'granularity' to get a day/week/month time series in one call instead of one flat total for the whole period.

get_gsc_keywords

Get complete Google Search Console report: period totals + top keywords + top pages, all properly aggregated.

get_gsc_full_report

Gets a single action/tiltak by id or key, with the full field set — everything list_actions returns plus description and the taxonomy/prioritization fields create_action and update_action can write (categoryId, workTypeId, tags, abcPhase, RICE fields, valueTypes, kpiIds, goalIds, planPointId, assigneeId, linkedUrlId). Use this when you need one action's complete record instead of paging through list_actions. `kpis` gives each linked KPI's id AND name together — refer to a KPI by its name in any prose/comment you write, never by the bare numeric kpiId.

get_action

Gets backlink monitoring dashboard statistics including total count, found/missing breakdown, and daily trend data

get_backlink_dashboard_stats

Gets a project's business profile data, including brand voice and visual identity

get_business_profile

One ranked overview per competitor for a period, merged from Web Spy, Sitemap Watch and Feed Watch — the call to make for a weekly competitor report instead of reading each module per agent. Each entry holds only what matters: Web Spy changes rated medium or above after grounding (minor ones are counted, not listed), new and removed sitemap URLs grouped by theme (the first path segment, e.g. `forbrukslan`, `refinansiering`), lastmod-only churn as a count, and new feed posts with their AI signal. `health` flags a competitor that appears to have shut down (Sitemap Watch lost the sitemap AND the site looks gone: a critical Web Spy change, every agent failing, or the feed failing) or whose monitoring keeps failing (e.g. an agent switched off after repeated errors), and lists the watches worth pausing. `openShutdownAlerts` lists competitors whose "appears to have shut down" alert is still open (it closes by itself, with a "back online" alert, once the sitemap — or, for a site whose pages had failed, the pages — can be read again). Ranked by `score` (critical change 10, high 6, medium 2; 2 per new URL theme, 1 per removed; launch/pricing post 4; a likely shutdown +100). Things watched without a competitor link are attributed by domain; anything still unmatched is its own entry named by host. Read-only; spends nothing.

get_competitor_weekly_digest

Gets the "us vs. them" Head-to-Head comparison for one competitor — AI citations/mentions, organic + paid traffic and keywords, topic coverage, and keyword overlap, each with a winner. The same computation the Head-to-Head page and its PDF export use, so an agent can read the comparison without the browser. Pass a competitor id from list_competitors.

get_competitor_head_to_head

Get complete index checker report for a project: all tracked URLs with their latest check results

get_index_checker_report

Reads the connection status of every marketing/analytics integration (Google Ads, GA4, Google Search Console, Meta Ads, GTM, WordPress, Contentful, Prepr, Sanity, Webflow, HubSpot, Umbraco) for a project in ONE call. Use this before calling an integration-specific tool (gtm.*, google_ads.*, ga4.*, search_console.*, cms.*) to check whether it is connected, instead of discovering 'not configured' one tool at a time. google_ads/ga4/gtm each carry `hasOAuthToken` separately from `connected`: `connected` means an id has been selected (customer/property/container), `hasOAuthToken` means this project has authorized that Google service at all. If `hasOAuthToken` is true but `connected` is false, connect_google_ads / connect_ga4 / connect_gtm can select an id without a new OAuth handshake; if `hasOAuthToken` is false, OAuth must be granted first via the Integrations page. Read-only.

get_integration_status

Keyword cannibalization from Google Search Console: queries for which two or more of the project's own URLs take impressions in the period, with each URL's clicks, impressions, average position and share of the query's impressions, the page that should own the query, the recommended method (merge, canonical, redirect, differentiate, or double_listing — two top-3 results, usually a gain) and the clicks the conflict costs. `sectionPages` lists competing pages other pages of the site sit under (the home page, a blog index): those are never merged or redirected — the method is differentiate, and the content page owns the query. Worst first. Pass 'keywords' to check specific queries; without it the property's top query×page rows are scanned. Spellings of one page (http/https, www, trailing slash, query string) count as one page. Brand searches and queries tagged brand or ignore on Analytics → GSC → Cannibalization are left out. Same rule as that tab (shared/cannibalization.ts, explore floor: a page counts from 2 impressions, and the competing pages need 5+ between them). Read live from the Search Console API. `competingSince` is the first day of the unbroken run of daily checks that saw the conflict (the stricter inbox floor on the stored 28-day window, recorded after each Search Console sync); null when the daily check has not seen it.

get_gsc_cannibalization

Gets details for a specific project by token, name, or domain

get_project

Gets a comprehensive summary of a project including counts of keywords, actions, and analyses

get_project_summary

Gets rank tracking keywords and their positions for a project, one device at a time (desktop by default; mobile only exists for keywords the project opted into mobile tracking for). Each keyword carries the SERP features present in its latest check (and whether the project owns each) and the positions of the project's competitors in that same SERP.

get_rank_tracking

Gets ranking position history for one tracked keyword over time. Identify the keyword EITHER by `keywordId` (from list_keywords) OR by `projectId` + `keyword` (the keyword text, matched case-insensitively among the project's tracked keywords) — the second form is what a diagnosis that names a search term can call directly.

get_ranking_history

Get a single recommendation with full context: business case (linked KPIs, keyword reach, URL scope, finding counts), linked actions, supporting findings (with source data), and event history. Use this when you need to understand the rationale, root cause, hypothesis, expected outcome, and quantitative anchors behind a rec before acting on it.

get_recommendation

Gets traffic referrer/source data for a project's website. Every row carries its channel (Organic Search, Paid Search, AI Assistant for ChatGPT / Perplexity / Gemini / Copilot / Claude, Referral, Direct, …) and channelTotals sums sessions per channel over the whole period, so AI traffic shows as its own channel.

get_referrers

Get issues found in a technical SEO audit. Optional filters (title, category, type, impact, pageId) are applied before the limit.

get_technical_seo_issues

Get complete technical SEO audit report: audit details + all pages + all issues + all links in one call

get_technical_seo_full_report

Gets the most visited pages for a project's website

get_top_pages

List AI Monitor prompts for a project. Per prompt, over ALL its completed runs (AI models and search engines alike): citationCount = citations of the project's own domains, mentionCount = mentions of the project's own brand, citationRate / mentionRate = share of completed runs (0-1) that cited / named the brand (null before the first completed run). For a date window use the visibility tools.

list_ai_prompts

Lists Google Ads campaigns for a project, one row per campaign per month (newest month first). Without filters it returns every campaign (including REMOVED) for every synced month, which can be very large — narrow it with status, yearMonth or startMonth/endMonth.

list_gads_campaigns

Lists KPIs for a project. `measurementTrust` is set when the Lab Brain diagnosis has the measurement behind a KPI in doubt (suspect/broken) — check that before acting on a move in it.

list_kpis

List Meta (Facebook/Instagram) Ads campaigns for a project — returns id, name, status, objective and budgets, as last synced from the connected Meta ad account.

list_meta_ads_campaigns

Lists comments on an action/tiltak

list_action_comments

Lists project actions with typed primary assignees. Filter assignee: cita for delegated work. agentReady is a deprecated read-only alias derived from Cita primary assignment. Done stays assigned; blocked runs wait with a report comment. `kpis` gives each linked KPI's id AND name together — refer to a KPI by its name in any prose/comment you write, never by the bare numeric kpiId.

list_actions

Lists all competitors for a project

list_competitors

Lists all projects in CitationLab with their details. Archived projects are left out unless includeArchived is true — every other tool refuses an archived project until it is restored in the app.

list_projects

List recommendations for a project with full business case enrichment. Each row includes linked KPIs (current/forecast/target/gap), keyword search-volume reach, URL scope (count + examples), source-finding counts, and persona count — whichever quantitative anchors are available on the rec. The response carries `total` (the full count of the filtered set) and `limit`/`offset` alongside `recommendations`, so a filtered set larger than one page is never mistaken for the complete set — page through with `offset` when `total` exceeds `offset + recommendations.length`.

list_recommendations