Quaro MCP server
Yes. Quaro publishes an official MCP server at https://mcp.quaro.io/mcp. It's hosted. It offers 206 tools. We last checked it on 22 Nov 2026.
Official: answers MCP requests at mcp.quaro.io, on Quaro's own domain quaro.io. Checked 22 Nov 2026. Maintained by Stuart Brameld.
Key facts
As of 22 Nov 2026, Quaro's official MCP server scores 55/100 (provisional) on MartechDB's Agent Readiness Index (v1.1). It answered 55 of our 55 checks in the last 90 days and lists 206 tools.
How the Agent Readiness Index works · Wrong? Suggest a correction
Quaro MCP server review
55/100
Quaro's MCP server scores 55/100 on MartechDB's Agent Readiness Index: usable by AI agents, with gaps.
The good
Hosted by Quaro: connect by URL, with nothing to install
Answered 100% of our checks in the last 90 days
The bad
Nothing we've found.
Not checked yet: how you sign in; a docs page listing its tools; what agent access costs. Anything not checked scores nothing.
Official MCP server
Quaro MCP server
- Address
- https://mcp.quaro.io/mcp
- Runs
- Hosted
- Sign-in
- Official MCP Registry
- Not listed
- How we know it's official
-
- Answers MCP requests at mcp.quaro.io, on Quaro's own domain quaro.io.
- Tools
- 206 tools · listed 22 Nov 2026
- First seen
- 1 Oct 2026
- Last checked
-
Reachable ·
16 Nov 2026: reachable17 Nov 2026: reachable18 Nov 2026: reachable19 Nov 2026: reachable20 Nov 2026: reachable21 Nov 2026: reachable22 Nov 2026: reachable
- Uptime
- 100% · 55 of 55 checks reachable since 1 Oct 2026
Quaro MCP server tools (206)
As the server lists them. Names and descriptions are the vendor's own.
-
delete_objectsDelete Objects[Quaro app chat] The app shows the user a confirm card with every value you send (name, location, language, device, interval, keywords ...) before this runs. That card IS the confirmation: do not ask for it in text and do not show a proposal table first. Where the text below says to confirm or never guess a setting with the user, use a sensible default for what the user did not state (location and language from the question or the language they write in) and call the tool now; ask only for values that cannot be derived (domain, brand, keywords, prompts). **Write operation - permanent.** Deletes SEVERAL objects of the customer's account in one call (use it whenever more than one object should be deleted, e.g. 'delete everything named X', so the user confirms the whole list once). For a single object, the specific delete tool is fine. Each item: {type, id | name, project_id, label}; label is required for every object addressed by id: the object's exact name as the list tool returned it,…
-
get_account_creditsGet Account CreditsRetrieve the credit overview for the authenticated account, one entry per suite (Research, Content, Monitoring, AI, and Quinn - the AI chat in the Quaro app, key `ki_chat_suite`). Every suite is always present; an `active` flag says whether the account has booked it — `active: false` means the rest of that suite's fields are absent, not that the lookup failed. With the personal key of a user who is not an account admin, the figures are shown only for the suites this user has access to; every other suite shows just its `active` flag (in the app the Credit Overview is visible to admins only). Does not consume credits. The Research Suite pool is spent by two things at once: ad-hoc lookups (keyword ideas, suggests, search volume, SEO analysis under /seo/*) and the monthly update of every keyword saved in a collection. `used` is the sum of both (`used_for_lookups` + `used_for_saved_collections`, broken out separately); this matches the number app.quaro.io's own Credit Overview shows, which…
-
get_account_meGet Account MeShows exactly what the API key you are using may do — which suites the account has booked, which of them this user can access, and with which role. Call it once before you start, or right after a 55021/55023/55024/55025 error, instead of finding out one denied call at a time. Does not consume credits. `key_type` is `"company"` for the legacy, company-wide API key (admin-equivalent) or `"user"`/`"mcp"` for a personal key. `user` is `null` for the legacy key — an API key that identifies an account, not a person, has no individual rights to report beyond the company's own suite bookings. For a personal key, `suites` has one entry per suite: `company_active` (the account has this suite booked — the same thing `the get_account_credits tool`' `active` flag reports), `user_access` (an admin has additionally given THIS user access to it — the reason behind a 55023), and `role` — `null` when either of the first two is false, otherwise one of `"full"`, `"author"` (Content Suite only), `"read_on…
-
list_account_invoicesList Account InvoicesLists every invoice, credit note and cancellation the account has received, newest first. Quaro uses two billing systems, and an account can have documents in both — this endpoint merges them into one list, so the caller never needs to know which system an account uses. `source` says which system each document comes from; `sources` says which billing systems are linked to this account at all (both `false` = no invoices yet, e.g. a trial account). Admin only: the company-wide API key and the personal API key (or MCP connection) of an account admin. Any other personal key gets 403 / status_code 55024. Does not consume credits. Amounts are in the invoice currency (not cents); `amount_net` excludes VAT, `amount_gross` includes it; both are reported as the billing system states them. `status` is only set for invoices (`paid`, `open`, `overdue` = open and past `due_date`, `canceled`); credit notes and cancellations have `status: null`. Drafts are never listed. PDFs: documents with `source`…
-
get_search_volume_statusGet Search Volume StatusRetrieve the latest available month for Quaro Search Volume and Google Search Volume data. Use it to check data freshness before requesting search volumes; both datasets update once a month. Does not consume credits.
-
list_google_searchvolume_locationsList Google Searchvolume LocationsRetrieve the available location codes for Google Search Volume. Use this endpoint to discover valid `location_code` values before requesting keyword metrics. Under `?response_mode=ai`, an unfiltered call returns only a 20-row preview (`meta.hint` explains this and points at `search`/`country`); pass `search` and/or `country` (optionally with `limit`) to look up specific entries instead of paging through the full list. Without `response_mode=ai`, the full list is always returned and `search`/`country`/`limit` have no effect. Does not consume credits.
-
get_google_search_volumeGet Google Search VolumeRetrieve search volume, CPC, SEA competition, and average search volume for the requested keywords. Rate limit: 2 requests per minute per API key — one budget shared with the get_keywords_for_domain tool and the get_keywords_for_url tool. Use `sv_length` to control how many months are included in the historical output. Difference to the get_quaro_search_volume tool: this endpoint looks up current Google Ads data for exactly the keywords you send, so every requested keyword gets a result. The Quaro variant instead reads from a monthly-updated, pre-fetched dataset (with history back to 2019) where a keyword can be missing if it was never monitored — prefer this endpoint when you need guaranteed coverage for specific keywords. Consumes Research Suite credits — 1 credit per call regardless of how many keywords are sent (credits are charged per request, not per keyword). `the interactive demo in the web docs` returns free static sample data in the same format. Personal keys need the Resear…
-
list_quaro_searchvolume_locationsList Quaro Searchvolume LocationsRetrieve the available location and language codes for Quaro Search Volume. Use this endpoint to discover valid `location_code` and `language_code` values before requesting keyword metrics. Does not consume credits.
-
get_quaro_search_volumeGet Quaro Search VolumeRetrieve search volume, CPC, SEA competition, categories, and average search volume for the specified keywords — plus, unlike the get_google_search_volume tool, a keyword difficulty score, SERP metadata and average backlink counts per keyword (`keyword_properties.keyword_difficulty`, `serp_info`, `avg_backlinks_info` in the response). The data is based on the Google Ads API. Keyword metrics are updated once a month. Historical metrics may be available from the beginning of 2019, depending on how long metrics have been available for a keyword. Personal keys need the Research Suite right id 11 (historical search volume); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted. Not every keyword is monitored in this dataset. If a keyword is not monitored, it can be missing in the response. If you need a guaranteed result for specific keywords, use the get_google_search_volume tool i…
-
get_keywords_for_domainGet Keywords For DomainRetrieve keywords, search volume, CPC, SEA competition, categories, and average search volume for a domain. Use `sv_length` (1-48 months, default 12) to set how many months are included in each keyword's `monthly_searches`. Use the get_keywords_for_url tool when you want the keywords of a single page instead of a whole domain. Consumes Research Suite credits — 1 credit per call regardless of how many keywords are returned (credits are charged per request, not per keyword). Optional `limit`/`skip` page the result. Omitting `limit` returns the full list. Rate limit: 2 requests per 60 seconds per API key — one budget shared with the get_google_search_volume tool and the get_keywords_for_url tool. A further call within the window is answered with a 429 that names the wait in `retry_after`. Personal keys need the Research Suite right id 8 (Ideas); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key…
-
list_monitoring_projectsList Monitoring ProjectsRetrieve all Quaro monitoring projects available in your Quaro workspace. The returned `_id` is the `project_id` every other monitoring endpoint expects, and `processed_dates` lists the crawl dates you can query. This route has no `skip`/`limit` of its own. `?response_mode=ai` caps the list to the first 30 projects, caps each project's `processed_dates` to the newest 24 entries, and drops `top_competitors`/`blacklisted_competitors` from each — reported via `meta.slimmed`. Use `response_mode=full` (the default) for the complete, uncapped list. The list also contains Quaro's shared example project (listed first, marked `read_only: true`): it is read-only (writes answer 403 / 55017) and its rankings, trends, clusters, keywords and AI Overviews only show an excerpt. Demo projects of your own account are listed like normal projects and carry `read_only: true` as well; writes to them are refused with 400 / 55017 (`additional_info`: "This is a demo project. It can be read but not changed.").…
-
list_monitoring_project_categoriesList Monitoring Project CategoriesRetrieve the categories configured for a Quaro monitoring project. Use `category_id` or `category_ids` to narrow the response and `sv_length` to limit monthly search-volume history. Without `category_id`/`category_ids` you have no way to know which categories exist before scoping a request — `skip`/`limit` let you page through the full list instead. Omitting both keeps the full, unsliced list. Every requested category id must exist on the project — an unknown one is rejected with 400 55016 (listed in `additional_info`) instead of being dropped; a non-string `category_id` (55016) or a `category_ids` that is not an array of strings (55029) is rejected too, as is a `skip`/`limit`/`sv_length` that is not an integer within its minimum (55029). `limit` values above 500 are capped at 500. Does not consume credits. `?response_mode=ai` caps the list to the first 100 categories and drops each category's `monthly_searches`/`categories` (the level-name map) — reported via `meta.slimmed`. Use `res…
-
create_monitoring_project_categoryCreate Monitoring Project Category**Write operation.** Creates a new category on a monitoring project, mirroring the "Add category" dialog in app.quaro.io. The new `category_id` is built from the category names joined with `;` (e.g. `"Shoes;Sneakers"` for a category named `"Sneakers"` under parent `"Shoes"`) — pass `parent_category_id` to nest under an existing category, or omit it for a top-level category. Creating an already-existing path is rejected rather than silently overwritten. Like in the app, creating a category only saves the category: it does NOT switch the project to categorized keywords (the app only does that in its setup wizard). To put keywords into the new category use the assign_monitoring_project_category tool. To switch the project to categorized keywords explicitly, send `categorized_project: true` to the update_monitoring_project tool. Only Monitoring projects are accepted; an AI Suite project_id is rejected with 55017. There is no dedicated rename route — rename a category by creating the new n…
-
assign_monitoring_project_categoryAssign Monitoring Project Category**Write operation.** Assigns, moves, or un-assigns the category of up to 1000 keywords in one call, mirroring the "assign to existing category" tab and the un-assign action in app.quaro.io. Pass `category_id` to assign/move the keywords into that category; omit it (or send null/empty string) to un-assign them back to uncategorized. Each keyword is matched as given and in its stored form (lowercased, special characters replaced — see the add_monitoring_project_keywords tool). The keyword counts of the affected categories are refreshed like in the app. Only Monitoring projects are accepted; an AI Suite project_id is rejected with 55017. Does not consume credits. Also possible while the first import of the project is still running, so the keywords of a project created with `categorized_project` set to true can be put into their categories right after the categories were created. A demo project in the account (an example project that is not updated and does not count towards the credits)…
-
delete_monitoring_project_categoryDelete Monitoring Project Category**Write operation.** Deletes a category (and every descendant category) from a monitoring project, mirroring the two-mode delete dialog in app.quaro.io. This cannot be undone. `mode: "delete"` removes the category and its keywords entirely; their reserved keyword updates are released at the end of the already-paid period, not immediately (see `pending_release` on the get_account_credits tool). `mode: "reassign"` keeps the keywords, moving them to `target_category_id` first — required only for this mode, and rejected with `mode: "delete"`. Either way, the deleted category (and its descendants) is gone afterwards. Only the category itself and its own subcategories are deleted (matched by full path) — a category with the same name under another parent is not affected. Only Monitoring projects are accepted; an AI Suite project_id is rejected with 55017. While the first import of the project is still running (its first crawl is not processed yet) the request is refused with 400 on `project…
-
get_monitoring_project_keywordsGet Monitoring Project KeywordsRetrieve keywords for a monitoring project. Supports pagination through `skip` and `limit`; a value that is not an integer within its minimum is rejected with 400 55029 (a `limit` above 10000 is capped). Like every monitoring route, an AI Suite project_id is rejected with 400 55017. In the shared demo monitoring project only the keywords its rankings show in full (the first five rows of the newest crawl in the default order) are returned with their data; every other entry is an object with only `blurred` set to true. `total` stays the full keyword count. Does not consume credits. Monitoring suite: the tracked ranking keywords of ONE monitoring project (position/visibility tracking). NOT a research keyword collection — for those use get_research_collection instead.
-
add_monitoring_project_keywordsAdd Monitoring Project Keywords**Write operation.** **Consumes Monitoring Suite keyword credits** only for keywords that are actually new: duplicates within the request and keywords already tracked in the project are free; per new keyword, updates per month (interval 1 → 30, 7 → 4, 14 → 2, 28 → 1), doubled when the project has ai_overview enabled. A paused project reserves nothing now (`reserved_keyword_updates: 0`, `paused: true`); the keywords are stored and the reservation is booked when the project is resumed. The response returns the reservation actually booked as `reserved_keyword_updates`. Add up to 1000 keywords to a Quaro monitoring project. Only Monitoring projects are accepted; an AI Suite project_id is rejected with 55017. Rate limit: 1 request per 10 seconds. While the first import of the project is still running (its first crawl is not processed yet) the request is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`) and nothing is changed; retry once the first crawl i…
-
delete_monitoring_project_keywordsDelete Monitoring Project Keywords**Write operation.** Delete keywords from a Quaro monitoring project. Each keyword is matched as given and in its stored form (lowercased, special characters replaced — see the add_monitoring_project_keywords tool), so "Running Shoes" removes the stored "running shoes". A submitted keyword that isn't actually tracked in the project is silently ignored rather than erroring — check `keywords_removed` against `keywords_submitted` in the response to tell whether everything you sent was actually removed. The keyword counts of the project's categories are refreshed like in the app. Only Monitoring projects are accepted; an AI Suite project_id is rejected with 55017. Rate limit: 1 request per 10 seconds. While the first import of the project is still running (its first crawl is not processed yet) the request is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`) and nothing is changed; retry once the first crawl is done. A demo project in the account (an exa…
-
get_monitoring_competitorsGet Monitoring CompetitorsRetrieve competitors for a monitoring project. Supports pagination and optional category filtering. `category_id` must be a category path of this project — a non-string or unknown value is rejected with 400 55016 instead of returning an empty list; a `skip`/`limit` that is not an integer within its minimum is rejected with 400 55029 (values above the maximum stay capped). Does not filter out hidden competitors — `hidden_competitors` (see the get_hidden_competitors tool) only hides domains in the tables of app.quaro.io; this endpoint always returns the full competitor list, hidden ones included. Does not consume credits.
-
search_monitoring_competitors_by_domainSearch Monitoring Competitors By DomainRetrieve competitor records for one or more domains in a monitoring project. `domains` is required and may contain up to 100 domains; every entry must be a string (otherwise 400 55029). `category_id` must be a category path of this project — a non-string or unknown value is rejected with 400 55016 instead of returning no rows; a `limit` that is not an integer of at least 1 is rejected with 400 55029. Does not consume credits.
-
get_top_monitoring_competitorsGet Top Monitoring CompetitorsRetrieve top competitors for a monitoring project, each with its two newest visibility entries. Supports category filters and paging with `skip`/`limit`; `total` is the number of competitors in the requested scope. `category_id` and every entry of `category_ids` must be category paths of this project — a non-string or unknown value is rejected with 400 (55016 for `category_id`, 55029 for `category_ids`, unknown entries listed in `additional_info`); a `skip`/`limit` that is not an integer within its minimum is rejected with 400 55029. Difference to the get_monitoring_competitors tool: this endpoint accepts `category_ids` to merge results across multiple categories, and returns a lighter record (no `_id`/`created`/`frequency` fields) — use `get` for the full record scoped to one category. Does not filter out hidden competitors — `hidden_competitors` (see the get_hidden_competitors tool) only hides domains in the tables of app.quaro.io; this endpoint always returns the full competitor li…
-
get_competitor_visibility_chartGet Competitor Visibility ChartMonitoring suite (tracked Google rankings): Retrieve visibility chart data for competitors. Supports optional domain filtering and category filtering; `skip` and `limit` page through the competitors in sort order (`limit` caps the number plotted per request) and `total` is the number of competitors matching the request. `domains` entries must be strings and `metric` a string from the enum (otherwise 400 55029); `category_id` must be a category path of this project — a non-string or unknown value is rejected with 400 55016 instead of returning an empty chart; a `skip`/`limit` that is not an integer within its minimum is rejected with 400 55029. Does not filter out hidden competitors — `hidden_competitors` (see the get_hidden_competitors tool) only hides domains in the tables of app.quaro.io; this endpoint always returns the full competitor list, hidden ones included. Does not consume credits. Chart-ready visibility series for the project's competitors, with domain/category filtering an…
-
discover_monitoring_competitorsDiscover Monitoring CompetitorsDiscover competitor changes between two dates. `date_current` and `date_prev` are required and must be two of the project's crawl days in `YYYY-MM-DD` format (see `processed_dates` in the list_monitoring_projects tool) — any other date is rejected with 400 55029 instead of returning an empty list. Use `category_id` for one category or `category_ids` for bulk mode across multiple categories; both must name categories of this project (unknown or non-string values are rejected with 400 — 55016 for `category_id`, 55029 for `category_ids`). `skip`/`limit`/`skip_per_category`/`limit_per_category` that are not integers within their minimum are rejected with 400 55029, as is a `sort` that is not one of the enum strings. **Two response shapes:** with a single category (or none) the response is a flat list under `competitors` with `total`/`skip`/`limit`. With two or more `category_ids` it switches to bulk mode and returns one entry per category under `categories`, paged with `limit_per_category…
-
get_competitor_domain_visibility_historyGet Competitor Domain Visibility HistoryRetrieve visibility history for one competitor domain. `domain` must contain exactly one domain. `category_id` must be a category of this project (a non-string or unknown value is rejected with 400 55016 instead of returning an empty history), and `months` an integer of at least 1 (otherwise 400 55029; values above 48 are capped). For multiple domains, use the search_monitoring_competitors_by_domain tool instead — note it returns the standard competitor-record shape (with top10/top50/top100 counts), not this endpoint's per-category `history` breakdown. Does not consume credits.
-
get_seller_shopping_visibilityGet Seller Shopping VisibilityRetrieve shopping visibility history for one or more seller names. Use `seller_name` or `seller_names`. **Required:** provide `seller_name` or `seller_names`. The names `domain`/`domains` are accepted as deprecated aliases for backwards compatibility; use `seller_names` in new integrations. A non-string `seller_name`/`domain` or a non-string entry in `seller_names`/`domains` is rejected with 400 55029, as is a `months` that is not an integer of at least 1 (values above 48 are capped). **Category:** `category_id` must be a category of this project (a non-string or unknown value is rejected with 400 55016). When no seller history is stored for that category, the whole project's history is returned instead and the response says so with `category_fallback: true` and a `hint`. Does not consume credits.
-
get_hidden_competitorsGet Hidden CompetitorsRetrieve the hidden competitor domains for a Quaro monitoring project. Like every monitoring route, an AI Suite project_id is rejected with 400 55017. Does not consume credits.
-
add_hidden_competitorsAdd Hidden Competitors**Write operation.** Hide one or more competitor domains for a Quaro monitoring project. Send either `domains` or `domain` in the JSON request body. Each domain is normalized to the bare host: lowercased; protocol, path, query, fragment and trailing '/' removed; a leading 'www.' is kept, because hidden domains are matched exactly against the competitor domain as listed by the get_monitoring_competitors tool. Domains already hidden are not added twice. **Required:** Provide `domain` or `domains`. While the first import of the project is still running (its first crawl is not processed yet) the request is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`) and nothing is changed; retry once the first crawl is done. A demo project in the account (an example project that is not updated and does not count towards the credits) can be read but not changed; a write is refused with 400 on `project_id` (status_code 55017, `additional_info` names the reason). A p…
-
delete_hidden_competitorsDelete Hidden Competitors**Write operation.** Remove one or more competitor domains from the hidden competitor list. Send either `domains` or `domain` in the JSON request body. Matching ignores protocol, path and a leading `www.`. Every entry of `domains` must be a domain string — a non-string entry (or a non-string `domain`) is rejected with 400 55029 ("Every entry of 'domains' must be a domain string."); an explicit empty `domains: []` is reported with attribute `domains`. **Required:** Provide `domain` or `domains`. While the first import of the project is still running (its first crawl is not processed yet) the request is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`) and nothing is changed; retry once the first crawl is done. A demo project in the account (an example project that is not updated and does not count towards the credits) can be read but not changed; a write is refused with 400 on `project_id` (status_code 55017, `additional_info` names the reason). A pe…
-
list_ai_projectsList Ai ProjectsRetrieve all Quaro AI tracking projects in your workspace, including per-project platforms, oldest tracked date per platform, the project brand and (optionally) the most-mentioned brands with the project visibility. `?response_mode=ai` caps the list to the first 30 projects and drops `top_competitors`/`ai_brand_colors`/`blacklisted_competitors` from each — use `skip`/`limit` to page through the rest, or `response_mode=full` (the default) for the complete, uncapped list. A demo project in the account is listed with `read_only: true` and can only be read through the API; a write to it is refused with 400 on `project_id` (status_code 55017). Does not consume credits.
-
get_ai_project_datesGet Ai Project DatesReturn the available AI-tracking dates for a project. With `platform` set, returns the earliest/latest/all dates for that platform; without it, returns the per-platform breakdown. `earliest`/`latest` always cover the full tracked span. `?response_mode=ai` caps the `dates` array(s) themselves to the newest 30 entries, reported via `meta.slimmed` — use `response_mode=full` (the default) for the complete list. A project whose first import is still running has no readable data yet - this and the other data endpoints of an AI project (brand visibility, prompts, mentions, brands, sources, fan-out queries, entities) answer 400 on `project_id` (status_code 55017, `additional_info` says that the project is still importing) until every tracked platform has its first run. Does not consume credits.
-
list_ai_project_tagsList Ai Project TagsList the tags (segments) configured on an AI tracking project. Defaults to the last two available dates; pass a date_scope to widen the window. Omit `platform` to aggregate tags across all platforms (the default window is then the full tracked range). This aggregates over prompt RESPONSES, not the tracked-prompt list itself — a tag set via `prompts` on the create_ai_project tool, the add_ai_prompts tool, or the set_ai_prompt_tags tool lands on the prompt immediately, but only appears here once that prompt has been crawled at least once (matches the general `import_status: "scheduled"` semantics of newly added prompts). Does not consume credits.
-
get_brands_ranking_single_dayGet Brands Ranking Single DaySingle-day brand ranking snapshot: every brand with its visibility, relevance rank, responses-with-mention, mentions, sources and sentiment for one day. Requires EXACTLY ONE day (a single day per platform when aggregating over all platforms). A tag filter narrows the list to brands actually mentioned under that tag. Does not consume credits. No native 'favorites only' filter exists. To restrict to the account's favorited brands (as in the app's Brands filter), first read `ai_brand_favorites` off the project from list_ai_projects, then pass that list as `filters.brands`.
-
get_brand_visibility_timeseriesGet Brand Visibility TimeseriesAI suite (GEO / AI-answer visibility), NOT Google-ranking visibility of a monitoring project (see get_project_visibility_history): Per-brand visibility time series over all available days. Requires filters.brands (the brands to chart). Omit `platform` to aggregate each brand's series across all platforms. **Series length:** with a `date_scope` each brand's `history` contains only the requested days; without one it is the complete recorded history. `sentiment_percent` is null for days before 2026-01-15. Does not consume credits.
-
get_top_brands_by_tagGet Top Brands By TagPer-tag brand breakdown: for each tag, the top brands by visibility on the selected day(s). Use filters.tags to restrict which tags are broken down and top_n to cap the brands per tag. Without `filters.tags` the request is limited to the first 25 tags of the project; `data.hint` reports the cap when it applies. Pass `filters.tags` to break down specific tags beyond that limit. Does not consume credits.
-
get_brand_performance_by_tagGet Brand Performance By TagTag × platform × brand matrix: for each selected brand, its visibility / relevance rank / sources / sentiment per tag and platform. Requires filters.brands. Use filters.platforms to limit platforms and filters.combine_with_tags to AND extra tags onto every analyzed tag. **Runtime:** one aggregation runs per tag and platform, so the response time grows with the requested date range — a single day answers in a few seconds, several months can take up to a minute. Narrow `date_scope` (and `filters`) when you need a fast answer. **Tags:** without `filters.tags` only the first 6 tags of the project are evaluated (the app's default) and `meta.hint` says so; pass `filters.tags` to choose others. A request may cover at most 100 tag × platform combinations - more is rejected with 400 55029 on `filters`. Does not consume credits.
-
get_brand_performance_by_platformGet Brand Performance By PlatformPer-platform brand metrics with a previous-vs-latest delta: for each platform, each brand's visibility / relevance rank / sources / sentiment and their change. Without filters.brands, the top brands per platform are used (cap with top_n). **Runtime:** like tag-performance, this aggregates per platform (and, for the delta, per period); wide date ranges can take noticeably longer to answer — narrow date_scope or filters.platforms for a fast answer. Does not consume credits.
-
get_ai_promptsGet Ai PromptsGet tracked prompts and their AI responses. They differ in the entity handling: mode=current filters the detected entities to the project's tracked brands, mode=all returns them unfiltered — mode=all is the default when `mode` is omitted. Both modes page the same way (1-500 rows per page, default 25). `total` reports how many rows match, so page through the set with `skip`. Does not consume credits. IMPORTANT: this data is per-platform and sparse — it often exists on only ONE platform (e.g. chatgpt) and only on that platform's own latest date. An all-platforms or unscoped query can return zero even when data exists. If you get an empty result, do NOT conclude there is none: call get_ai_project_dates, then query this tool once PER platform using that platform's latest date.
-
get_brand_mentionsGet Brand MentionsGet the brand mentions behind the AI responses. mode=current applies the dashboard's entity filtering; mode=all does not and is the default when `mode` is omitted — both page the same way (1-500 rows per page, default 25); mode=new/lost compare the first and the last day of the range and return the diff in one call (no true pagination); each of the two days is read up to 5000 answers, and `truncated` is true when a day reached that limit, so the diff may be incomplete. All modes report the match count in `total`. Does not consume credits. IMPORTANT: this data is per-platform and sparse — it often exists on only ONE platform (e.g. chatgpt) and only on that platform's own latest date. An all-platforms or unscoped query can return zero even when data exists. If you get an empty result, do NOT conclude there is none: call get_ai_project_dates, then query this tool once PER platform using that platform's latest date. mode=new/lost additionally REQUIRES an explicit `date_scope` — omitting i…
-
get_brands_snapshot_latest_dayGet Brands Snapshot Latest DayGet the brands table (visibility, mentions, sources, sentiment) for the latest day, with a previous-vs-latest delta when the date scope contains two days. The `has_comparison` response field flags whether deltas are present. Does not consume credits. No native 'favorites only' filter exists. To restrict to the account's favorited brands (as in the app's Brands filter), first read `ai_brand_favorites` off the project from list_ai_projects, then pass that list as `filters.brands`.
-
get_cited_sourcesGet Cited SourcesGet the cited sources for the latest day. mode=domain aggregates by domain; mode=url lists individual URLs. A previous-vs-latest delta is included when the date scope contains two distinct days (single platform). Does not consume credits. IMPORTANT: this data is per-platform and sparse — it often exists on only ONE platform (e.g. chatgpt) and only on that platform's own latest date. An all-platforms or unscoped query can return zero even when data exists. If you get an empty result, do NOT conclude there is none: call get_ai_project_dates, then query this tool once PER platform using that platform's latest date.
-
get_fanout_queriesGet Fanout QueriesGet the fan-out queries the AI platforms generated for the tracked prompts. mode=current filters entities to the tracked brands; mode=all does not and is the default when `mode` is omitted — both page the same way (1-500 rows per page, default 25). `total` reports how many rows match — see the `total` field description for a caveat about fan-out query filtering. Does not consume credits. IMPORTANT: this data is per-platform and sparse — it often exists on only ONE platform (e.g. chatgpt) and only on that platform's own latest date. An all-platforms or unscoped query can return zero even when data exists. If you get an empty result, do NOT conclude there is none: call get_ai_project_dates, then query this tool once PER platform using that platform's latest date.
-
get_keyword_ideasGet Keyword IdeasReturns keyword ideas for one or more seed keywords from a Google Ads-based keyword database, including search volume, CPC, competition and SERP info. This is the server-side equivalent of the "Ideas → by keyword" view in app.quaro.io. Defaults to 1000 ideas when `limit` is omitted; send `limit` (max 5000) to fetch more, or a smaller value for a smaller page. For live autocomplete-based suggestions instead, use the get_autocomplete_suggestions tool. Consumes Research Suite credits — 1 credit per call regardless of how many seed keywords are sent or ideas are returned (credits are charged per request, not per keyword). `?response_mode=ai` additionally caps the response to the first 150 rows (each row's `monthly_searches` capped to 12 months, `keyword_annotations` dropped), reported via `meta.slimmed` — independent of and on top of the `limit` parameter above. Use `response_mode=full` (the default) for the complete result. Personal keys need the Research Suite right id 8 (Ideas); withou…
-
get_keyword_difficultyGet Keyword DifficultyReturns the keyword difficulty (0–100) for a list of keywords. The score estimates how hard it is to rank in the organic top 10 for the given location. Consumes Research Suite credits — 1 credit for up to 650 keywords, 2 credits above; `meta.credits_consumed` reports the charge. Personal keys need the Research Suite right id 10 (Google search volume); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
get_search_intentGet Search IntentClassifies the search intent (informational, navigational, commercial, transactional) of a list of keywords, including the probability per intent. This endpoint is language driven; `location_code` is optional. Consumes Research Suite credits — 1 credit per call regardless of how many keywords are sent. Personal keys need the Research Suite right id 10 (Google search volume); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
get_autocomplete_suggestionsGet Autocomplete SuggestionsCollect keyword suggestions the way the Suggests view in the app does: the seed keyword is expanded (seed, seed+letter, letter+seed, seed+digit, digit+seed) and every variation is looked up against the platform's live autocomplete service. **Platforms:** `google`, `questions` (w-questions: each question/preposition/comparison word is asked before and after the seed), `images`, `shopping`, `youtube`, `news`, `amazon` (17 marketplaces), `ebay` (16 shops) and `bing`. **Reliability:** the response reports how the lookups went. `requests_direct` and `requests_proxied` count the lookups that were answered (directly, or through a fallback route), `requests_failed` those that stayed unanswered even after one retry — a value above zero means the keyword list may be incomplete. If not a single lookup answers, the endpoint returns 502 instead of an empty result — nothing is charged in that case. **Runtime:** a full expansion is 40–80 autocomplete lookups. The endpoint works within a hard 60-seco…
-
list_research_collectionsList Research CollectionsLists all keyword collections of the account with keyword count, location, language, status and the monthly search-volume trend. Use the returned `name` as the `collection` attribute of the other collection endpoints. Does not consume credits. Personal keys need the Research Suite right id 1 (see collections and research projects) or right id 7 (add keywords); without one of them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
list_research_collection_categoriesList Research Collection CategoriesLists the research categories of the account, including the custom categories created in app.quaro.io. Categories are hierarchical: `category_code_parent` points at the parent category. Under `?response_mode=ai`, an unfiltered call returns only a 20-row preview; pass `search` (optionally with `limit`) to look up specific categories instead. Without `response_mode=ai`, the full list is always returned and `search`/`limit` have no effect. Does not consume credits.
-
create_research_collectionCreate Research Collection**Write operation.** Creates a new keyword collection and enriches every keyword with historic search volume, CPC, competition, keyword difficulty and SERP item types — exactly like creating a collection in app.quaro.io. Consumes Research Suite credits: saved keywords count against the budget (one credit per 700 saved keywords). When the credits left do not cover the new keywords, the request fails with status_code 50000. The collection name must be unique for the account (case-insensitive). This request can run up to 180 seconds. Personal keys need the Research Suite right id 6 (create collections); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted. Creates a saved keyword collection exactly like the app does: the keywords are enriched with 48 months of search volume, CPC, difficulty and SERP features (spends research credits, takes ~10-30s). Verify afterwards with get_res…
-
delete_research_collectionDelete Research Collection**Write operation.** Permanently deletes a keyword collection together with all of its keywords. This cannot be undone. `deleted: true` is returned only after the collection and all its keywords have been removed. If the deletion fails, the response is 500 with `status_code` 56000, and the collection may still exist (possibly without keywords) — retry the delete. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additional_info` says how many seconds ago the import last reported progress. Personal keys need the Research Suite right id 1 (see collections and research projects) plus either right id 3 (delete any collection) or right id 28 (delete own collections, only as the owner of this collection); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_accou…
-
get_project_rankingsGet Project RankingsMonitoring suite (tracked Google rankings): Returns the keyword ranking table of a monitoring project for a single crawl date: one row per keyword with search volume, the tracked domains' positions and the SERP features found. This is the data behind the project's rankings table in app.quaro.io. **Note on `amazon_search_volume`:** the value is the **monthly** Amazon search volume. The app's "Amazon searches (yearly)" column multiplies it by 12 for display, so an API value of 444 appears as 5,328 in the UI. **Internal search volume:** rows carry `internal_search_volume` (newest upload) and `internal_search_volume_prev` (previous upload) whenever the account has an internal site-search project whose domain matches this monitoring project. This mirrors what the app's rankings table shows; without such a project the fields are absent. **Crawl dates:** pass the plain day (`YYYY-MM-DD`) exactly as listed in a project's `processed_dates`. Crawls that ran at a specific time of day are resolve…
-
get_project_keyword_trendsGet Project Keyword TrendsMonitoring suite (tracked Google rankings): Returns the keyword trend table of a monitoring project: one row per keyword with its search-demand development — monthly Google searches (`ads_data.monthly_searches`), short and long moving averages, and the month and year change (percent and absolute). It contains no ranking positions; use the get_project_rankings tool for the positions on a crawl date. **Note on `amazon_search_volume`:** the value is the **monthly** Amazon search volume. The app's "Amazon searches (yearly)" column multiplies it by 12 for display, so an API value of 444 appears as 5,328 in the UI. **Validation:** `category_id` must be a category of this project (a non-string or unknown value is rejected with 400 55016 instead of returning an empty table); `search` must be a string and `skip`/`limit` integers within their minimum (a `limit` above 500 is capped), otherwise 400 55029. Does not consume credits. **Shared demo project:** in the shared demo monitoring project onl…
-
get_rankings_for_keywordsGet Rankings For KeywordsReturns the stored rankings of a domain for a list of keywords across ALL monitoring projects of the account — no project_id required. Only keywords that are tracked in a monitoring project with the same `domain`/`location_code`/`device` can resolve; unmatched keywords are simply absent from the result. `domain` is always matched in its "www." form (see the `domain` field). A project whose tracked domain was stored without "www." (e.g. "example.com") is not matched by this endpoint, as in app.quaro.io — use the get_project_rankings tool with its `project_id` instead. Each resolved row's `ranking` sub-object belongs to the requested `domain`, and is absent for a keyword where that domain doesn't rank in the tracked project. `keywords` must be an array of strings, `location_code` a positive integer and `device` a string — anything else is rejected with 400 55029. Does not consume credits.
-
compare_competitor_visibilityCompare Competitor VisibilityReturns the competitor comparison table of a monitoring project: one row per competitor domain with its visibility on the selected crawl dates, so the change between two dates can be evaluated. In the returned `visibility` array index 0 is the current date and index 1 the comparison date. **Paging:** this endpoint always returns the full competitor set for the project; `limit` is not supported and is rejected. Use `skip` to offset or `search` to narrow the result. **Validation:** the dates are `date_current`/`date_prev` (a `dates` array is rejected with 400 55029) and must be crawl days of the project — a day without a crawl is rejected with 400 55029 instead of returning empty `visibility` arrays. `category_id` must be a category of this project (a non-string or unknown value is rejected with 400 55016); `search` must be a string and `skip` an integer of at least 0 (otherwise 400 55029). Difference to the discover_monitoring_competitors tool: `discovery` is paged, sortable by change,…
-
get_project_category_breakdownGet Project Category BreakdownMonitoring suite (tracked Google rankings): Returns the category breakdown of a monitoring project: for every category on the requested nesting level, the visibility and ranking data of the selected competitor domains. This is the data behind the category comparison view in app.quaro.io. **Crawl dates:** pass the plain day (`YYYY-MM-DD`) exactly as listed in a project's `processed_dates`. Crawls that ran at a specific time of day are resolved to that stored timestamp automatically, so every advertised crawl date returns data. If a day holds more than one crawl, the most recent one is used. **Freshness:** like the get_category_search_volume_trends tool, this reads aggregates stored per crawl run rather than recomputing live on every request — values reflect the state as of each requested crawl date, not the current moment. **Validation:** every entry of `dates` must be one of the project's crawl days and `domains` an array of strings, `level` an integer 0-10 and `filters` an object — a…
-
get_keyword_ai_overviewGet Keyword Ai OverviewMonitoring suite (Google SERP feature, NOT the AI suite): Returns the Google AI Overview (AIO) recorded for one keyword of a monitoring project on a given crawl date, including the answer text and the cited source links. **Not to be confused with the AI Suite:** this is a classic Monitoring Suite SERP feature — the AI Overview box shown on a regular Google search results page for one tracked keyword. It is unrelated to the AI Suite's `google_ai_overview` platform, which tracks brand mentions across AI answers for a set of tracked prompts (see the list_ai_projects tool). **Crawl dates:** pass the plain day (`YYYY-MM-DD`) exactly as listed in a project's `processed_dates`. Crawls that ran at a specific time of day are resolved to that stored timestamp automatically, so every advertised crawl date returns data. If a day holds more than one crawl, the most recent one is used. A well-formed day without a crawl is rejected with 400 55029 instead of returning an empty result. Only this proje…
-
update_monitoring_folderUpdate Monitoring Folder**Write operation.** Renames a folder and/or replaces the list of monitoring projects assigned to it. Attributes that are omitted stay unchanged (a rename keeps the assigned projects); passing `project_ids` REPLACES the whole assignment list. At least one of `name`/`project_ids` is required. An unknown `folder_id` is rejected with 400 55029 (attribute `folder_id`). A personal key with a read-only Monitoring Suite role gets 403 / 55024. Does not consume credits.
-
delete_monitoring_folderDelete Monitoring Folder**Write operation.** Deletes a folder. The monitoring projects that were assigned to it are not deleted, they only lose the folder assignment. This cannot be undone. An unknown `folder_id` is rejected with 400 55029 (attribute `folder_id`) instead of being answered `deleted: true`. A personal key with a read-only Monitoring Suite role gets 403 / 55024. Does not consume credits.
-
get_all_brands_kpis_rangeGet All Brands Kpis RangeReturns the aggregated brand KPIs (visibility, mentions, sources, sentiment, rank) of an AI tracking project over an explicit day window, optionally compared against a second window. Unlike the snapshot endpoints this always takes a `{from,to}` range rather than a list of crawl days. When `comparison_scope` is omitted, the equally long window ending right before `date_scope.from` is used for the comparison. **Runtime and range limit:** the calculation covers every AI answer and cited source in the window, so response time grows with the requested date range. Both `date_scope` and `comparison_scope` are limited to 31 days each (so any calendar month fits) — a wider request is rejected with a 400 rather than answered slowly. Narrow `filters` for a faster answer within that limit. Does not consume credits. The AI-visibility dashboard KPIs for ALL brands at once (visibility %, sentiment, source references, with previous-period deltas), over an explicit {from,to} range. For a single latest…
-
get_single_brand_kpis_rangeGet Single Brand Kpis RangeReturns the detailed KPI metrics (visibility, sentiment, rank, source references) of ONE brand in an AI tracking project over an explicit day window, optionally compared against a second window. **`brand` is validated against the project:** it must match the project's own brand or one of its `top_brands` (including their canonical-name variants and mention aliases) — see the get_brands_snapshot_latest_day tool for the tracked set. An unrecognized brand is rejected with a 400 listing up to 5 known brand names, rather than a misleadingly empty "0 visibility / no sentiment" success response. **Runtime and range limit:** the calculation covers every AI answer and cited source in the window, so response time grows with the requested date range. Both `date_scope` and `comparison_scope` are limited to 31 days each (so any calendar month fits) — a wider request is rejected with a 400 rather than answered slowly. Narrow `filters` for a faster answer within that limit. Does not consume credits.…
-
search_project_entitiesSearch Project EntitiesSearches the entities (brands, products, organisations, concepts) that were recognised in the AI answers of a project. Use it to find out which entities the models associate with your topics. **Runtime:** the search scans the project's full response corpus, so larger projects can take longer to answer (up to roughly a minute). This endpoint has no date or filter scoping — narrow `query` or adjust `limit` if you need a faster or more targeted answer. Not available for the read-only AI Suite role (403, status_code 55024) - in the app this search is part of managing the brands of a project. Does not consume credits. Full-text search over the entities extracted from a project's AI answers, with occurrence counts. Use suggest_entities (Knowledge Graph) to resolve a brand name BEFORE creating prompts or brands. Not available for a read-only AI Suite role (403) - in the app this search is part of managing a project's brands.
-
suggest_entitiesSuggest EntitiesKnowledge-graph lookup, not project data: resolves a free-text search term to canonical entity names. Use it to find the exact canonical brand name before creating brands or prompts in an AI tracking project. Does not consume credits. Not available for the read-only AI Suite role (403, status_code 55024) - the suggestions belong to setting up a project. The rate limit of this endpoint is 30 requests per 60 seconds.
-
get_response_entitiesGet Response EntitiesReturns the entities that were recognised in ONE recorded AI answer. Use the `id` of a response returned by the get_prompt_overview tool. Does not consume credits.
-
get_prompt_overviewGet Prompt OverviewReturns everything recorded for a single tracked prompt: the summary KPIs, the time series, the per-platform breakdown, the top cited sources, the recognised entities and the most recent AI answers. Identify the prompt either by its text (`prompt`) or by its `task_id`; at least one of the two is required. Omitting `date_scope` returns the full recorded history of the prompt. **Required:** provide either `prompt` (the exact prompt text) or `task_id`. A request with neither is rejected with status_code 55029. **Runtime:** this aggregates every recorded response for the prompt across six sections; omitting `date_scope` on a long-tracked, high-volume prompt can take noticeably longer than a narrow query — pass `date_scope` to bound it. Does not consume credits. Deep-dive into a single tracked prompt: response count, per-platform timeseries, top sources, entities and recent AI answers. Identify the prompt via `prompt` (exact text) or `task_id` — both come from get_ai_prompts.
-
list_content_projectsList Content ProjectsLists the content projects of the account with their main keyword, content score, word count, workflow status and folder assignment. The list includes the values the filters work on (`device`, `current_stage`, `authors`). Use the returned `project_id` for the other Content Suite endpoints. With a personal API key the list follows the app's visibility rules — a user with the Content Suite role Author only gets the projects they are listed on as an author, and a project whose setup is not finished is only listed for the user who created it. Admins see every project, users with full Content Suite access see every finished project, and the account-wide key is not restricted. `?response_mode=ai` returns at most the first 100 projects (reported in `meta`). Does not consume credits.
-
list_content_foldersList Content FoldersLists the folders used to organise content projects, together with the IDs of the content projects assigned to each folder. Does not consume credits.
-
create_content_folderCreate Content Folder**Write operation.** Creates a folder to organise content projects, optionally assigning existing content projects to it right away. The folder name must be unique for the account (case-insensitive). Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted.
-
update_content_folderUpdate Content Folder**Write operation.** Renames a folder and/or changes which content projects are assigned to it. Only the sent fields are changed; at least one of `name`/`project_ids` is required (a request with neither is rejected with a 400, attribute `fields`). Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted.
-
delete_content_folderDelete Content Folder**Write operation.** Deletes a folder. The content projects that were assigned to it are NOT deleted — they simply become unassigned. Irreversible. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted.
-
get_content_projectGet Content ProjectReturns the full document of one content project, including its settings, setup state, competitor configuration and derived status. Internal processing data and technical fields of the underlying keyword and SERP data are not part of the response; `authors` lists the authors by user name only, and `workflow` only carries the title of the active stage (see the get_content_project_workflow tool for the stages and tasks). `?response_mode=ai` caps `competitors` to the top 10 and `topics` and `questions` to the first 50 each, caps each entry of `topics[].keywords` (the keyword cluster per topic) to the first 10, and drops the `keyword_info`/`serp_info` sub-fields of every topic and the `wdfidf_data`/`lighthouse` sub-fields from both `competitors` entries and `comparissonURL` (a very large field on the full response) — reported via `meta.omitted`. Use `response_mode=full` (the default) for the complete, uncapped document. With a personal API key the project has to be visible to the user of…
-
get_content_project_keywordsGet Content Project KeywordsReturns the stored WDF*IDF keyword result of a content project: per keyword the WDF value, the IDF value, the coverage against the competitors and the term placement in the editor text. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects tool); otherwise the call answers 400 as for an unknown `project_id`. `?response_mode=ai` returns at most the first 100 keywords and drops the per-URL `wdf` array of every keyword (the aggregates such as `avgWDF`, `maxwdf`, `avg_wdf_idf` and `idf` stay); what was reduced is reported in `meta`. Does not consume credits.
-
get_content_project_editorGet Content Project EditorReturns the editor document of a content project: the current HTML content together with the content score, word count and the timestamp of the last update. `locked` tells you whether the editor is currently open in app.quaro.io (someone was active in it within the last 15 seconds) — check it before calling the save_content_project_editor tool to avoid overwriting someone who is typing right now. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects tool); otherwise the call answers 400 as for an unknown `project_id`. Does not consume credits.
-
get_content_project_briefingGet Content Project BriefingReturns the briefing description stored on a content project — the free-text brief the editor works against. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects tool); otherwise the call answers 400 as for an unknown `project_id`. Does not consume credits.
-
get_content_project_workflowGet Content Project WorkflowReturns the workflow of a content project: its stages, each with tasks that carry assignees, due dates, completion state and comments. `workflow` is null when the project has no workflow yet — create one with the create_content_workflow tool. People are always identified by email (see the list_content_suite_users tool for the addressable set), never by an internal id. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects tool); otherwise the call answers 400 as for an unknown `project_id`. Does not consume credits.
-
list_monitoring_foldersList Monitoring FoldersLists the folders used to organise the monitoring projects of the account, together with the project IDs assigned to each folder. Folders belong to the whole account, so no `project_id` is required. Does not consume credits.
-
create_monitoring_folderCreate Monitoring Folder**Write operation.** Creates a new folder for organising monitoring projects and optionally assigns projects to it. The folder name must be unique for the account (case- and accent-insensitive). A personal key with a read-only Monitoring Suite role gets 403 / 55024. Does not consume credits.
-
list_project_keyword_clustersList Project Keyword ClustersMonitoring suite (tracked Google rankings): Returns the keyword clusters of a monitoring project. Clusters group keywords that share a common SERP intent. Pass `parent_cluster_id` to walk into the sub-clusters of a cluster; omit it for the top level. Responses can get very large (a single cluster can carry hundreds of keywords and domain shares) — with `?response_mode=ai`, each cluster's `keywords` and `children` are capped to the first 10 and `domain_share` to the top 8 by share, with the drops listed in `meta.omitted`; the default `full` mode returns everything uncapped. `parent_cluster_id` must be a 24-character hex cluster id, `search` a string and `skip`/`limit` integers within their minimum (otherwise 400 55029); `category_id` must be a category of this project (a non-string or unknown value is rejected with 400 55016). Does not consume credits. In the shared demo monitoring project only the first five clusters are returned in full; every further cluster is blurred (`blurred` se…
-
get_project_keyword_clusterGet Project Keyword ClusterMonitoring suite (tracked Google rankings): Returns one keyword cluster of a monitoring project with its full detail, including the keywords assigned to it. Does not consume credits. The cluster has to belong to the given project - the id of a cluster of another project of the account is rejected with 400 55017 on `project_id`.
-
get_research_collectionGet Research CollectionReturns the keyword rows of one collection with average search volume, monthly history, CPC, competition, keyword difficulty and the assigned categories. Rows are sorted by average search volume (descending) and can be filtered and paginated server-side. Does not consume credits. Personal keys need the Research Suite right id 1 (see collections and research projects); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted. NOTE: `collection` is the saved collection's NAME (not an id) — discover the names with list_research_collections.
-
get_project_serp_featuresGet Project Serp FeaturesMonitoring suite (tracked Google rankings): Two-step endpoint. Call it without `serp_feature` to get `available_serp_features` — the SERP feature types recorded for the project (e.g. featured_snippet, people_also_ask, popular_products, ai_overview); `frequencies` is empty in that case. Call it again with one of these values as `serp_feature` to get how often that feature occurred on each crawl date. A non-string `serp_feature` is rejected with 400 55029; `category_id` must be a category of this project (a non-string or unknown value is rejected with 400 55016). Does not consume credits.
-
get_project_shopping_sellersGet Project Shopping SellersMonitoring suite (tracked Google rankings): Returns the sellers found in a shopping-style SERP feature (default `popular_products`) of a monitoring project. **Two modes:** without `seller_names` the endpoint returns a snapshot of all sellers present on a single crawl date (`date`, default: latest crawl). With `seller_names` it returns the per-date history of exactly those sellers across all crawl dates; combining `seller_names` with `date` is rejected with a 400. **Crawl dates:** pass the plain day (`YYYY-MM-DD`) exactly as listed in a project's `processed_dates`. Crawls that ran at a specific time of day are resolved to that stored timestamp automatically, so every advertised crawl date returns data. If a day holds more than one crawl, the most recent one is used. A well-formed day without a crawl is rejected with 400 55029. **Validation:** `seller_names` must be an array of strings and `serp_feature` a string (otherwise 400 55029); `category_id` must be a category of this project (a…
-
get_source_page_typesGet Source Page TypesReturns how the sources cited by the AI answers of a project split across page types (e.g. blog, product, forum, news). Intended for a narrow window; when `platform` is omitted the day(s) are resolved per platform and echoed back as `date_scope.by_platform`. `date_scope` accepts a single day or a multi-day range/list and aggregates across it, but is capped like the app's view — one day when `platform` is given, two days per platform when `platform` is omitted — because every cited source URL in the window has to be classified. A wider `date_scope` is rejected with a 400 rather than answered slowly. Only `project_id`, `platform`, `date_scope` and `filters` are accepted; any other top-level field (e.g. `sort`, `skip`, `limit`, `mode`, `top_n`, `total_responses`, `show_top_brands`) is rejected with a 400 on `request`. Does not consume credits. Which page TYPES (blog, product page, docs, video, ...) the AI answers cite. `total_urls` is the sum over all types.
-
run_wdfidf_analysisRun Wdfidf Analysis**Write operation.** Runs a full ad-hoc WDF*IDF analysis for a keyword: crawls the live Google SERP for the given location and device, scrapes the top ranking competitor pages and computes the WDF*IDF term profile. **Consumes one Content Suite WDF*IDF credit** per successful run (`content_suite.adhoc_wdfidf` in the get_account_credits tool), echoed as `meta.credits_consumed`. The credit is reserved when the call starts and given back when the analysis fails (502), so parallel calls cannot exceed the limit. Location and language have to be one of the nine pairs the app offers (see `location_code`). `?response_mode=ai` returns at most the first 100 terms without their per-URL `wdf` array and the first 10 competitors without their `wdfidf_data`/`lighthouse` sub-fields (reported in `meta`). This request is long-running and can take up to 160 seconds. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me…
-
get_keywords_for_urlGet Keywords For UrlRetrieve the keywords a single URL ranks for, with search volume, CPC, SEA competition and monthly search-volume history — the URL-level counterpart of the get_keywords_for_domain tool. Use this when you want the keywords of one specific page (for example a category or article page). Use the domains endpoint when you want the keywords of a whole domain. Consumes Research Suite credits — 1 credit per call regardless of how many keywords are returned (credits are charged per request, not per keyword). `data` is an array, one object per keyword. Optional `limit`/`skip` page the result. Omitting `limit` returns the full list. Rate limit: 2 requests per 60 seconds per API key — one budget shared with the get_google_search_volume tool and the get_keywords_for_domain tool. A further call within the window is answered with a 429 that names the wait in `retry_after`. Personal keys need the Research Suite right id 8 (Ideas); without it the call is refused with 403 / status_code 55024 — see `res…
-
add_collection_keywordsAdd Collection Keywords**Write operation.** Adds keywords to an existing collection and enriches each new keyword with historic search volume, CPC, competition and SERP item types — the API equivalent of pasting keywords into an open collection in app.quaro.io. Location and language are **taken from the collection**, not from the request, so the added keywords are measured exactly like the ones already stored. A very old collection without a stored location cannot be extended and answers 400. Keywords that are already in the collection are skipped (case-insensitive) and reported as `skipped_existing`; only genuinely new keywords are enriched and billed. Duplicates within the request (case-insensitive) are collapsed first and counted in `skipped_existing` too. At most 700 keywords per request after de-duplication; each keyword at most 80 characters and 10 words. The returned `keyword_count` is the recounted total of the collection. Consumes Research Suite credits: saved keywords count against the budget (one…
-
delete_collection_keywordsDelete Collection Keywords**Write operation.** Removes the named keywords from a collection and corrects the collection's keyword count. The collection itself is kept — use the delete_research_collection tool to remove the whole collection. Keywords that are not in the collection are ignored. `keywords` reports how many keywords you sent, `keywords_matched` how many of them existed in the collection and were removed. Does not consume credits. This cannot be undone. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 5 (remove keywords); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
categorize_collection_keywordsCategorize Collection Keywords**Write operation.** Assigns one category to the given keywords of a collection and makes it the active (displayed) category — the API equivalent of setting the category in the Categories column of a collection table. The category is identified by its numeric `category_code`. Get the valid codes for the account from the list_research_collection_categories tool. Quaro resolves the full category path (parent chain) itself, so you only pass the leaf code. The same call works whether or not the keyword had a category before: an existing category is switched, a missing one is added and activated. **Verifying the result:** read the keywords back with the get_research_collection tool. That endpoint returns the assigned categories. the legacy raw keyword export (not available as a tool) returns the raw categories from the data source instead, so it will not show your assignment. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list…
-
set_collection_column_valueSet Collection Column Value**Write operation.** Writes the same value into a custom column for the given keywords of a collection — the API equivalent of filling a custom column cell in the collection table. The column must already exist — create it with the create_collection_column tool and discover the defined columns with the list_collection_columns tool. The value is validated against the column type: a `dropdown` column only accepts one of its option texts (the rejection lists the allowed values), a `number` column only accepts numeric values, a `text` column accepts any text. Sending an empty value clears the cell. Keywords that are not in the collection are skipped silently — compare `keywords_matched` with the number of keywords you sent to detect typos. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additional_info…
-
get_category_search_volume_trendsGet Category Search Volume TrendsMonitoring suite (tracked Google rankings): Returns the categories of a monitoring project with their **search-volume development aggregated from the project's keywords**, plus the change metrics shown in the category charts of app.quaro.io. Difference to the list_monitoring_project_categories tool: both endpoints return the same stored per-category figures (`monthly_searches` aggregated from the category's keywords, plus the change metrics). This endpoint returns the rows in the order of the requested `category_ids` and adds `project_id` to each row. Use either one for market-development questions ("is the demand in this category growing?"). **Freshness:** `monthly_searches` is recomputed once per crawl run, not live on every request — it reflects the keywords assigned to each category as of the most recent crawl and can go stale between crawls. Use `sv_length` to keep the response small — without it, the full history is returned per category. Without `category_id`/`category_ids` you…
-
get_project_visibility_by_domainGet Project Visibility By DomainMonitoring suite (Google visibility, NOT AI-answer visibility): Returns the visibility index of every domain tracked in a monitoring project — your own domain and its competitors — with one data point per crawl date, including the ranking distribution (top 10 / 50 / 100) and how often the domain was cited in Google AI Overviews. This is the data behind the visibility chart of a project. Use it to answer "how visible is my domain in this market, and how does that compare to the competition?". Difference to the get_monitoring_competitors tool: that endpoint is the competitor table with its filters and sorting. This endpoint returns the plain visibility series per domain and is the smaller response when you only need the curves. For a chart-ready time series (one selectable metric, transposed by date) instead of a per-domain list, see the get_competitor_visibility_chart tool. Does not filter out hidden competitors — `hidden_competitors` (see the get_hidden_competitors tool) only hides do…
-
get_project_visibility_historyGet Project Visibility HistoryMonitoring suite (Google visibility, NOT AI-answer visibility): Returns the **visibility of the project over time** — one value per crawl date, summed over all stored visibility entries of the project (every domain, project-wide and per-category entries alike; at most 1000 entries are read), optionally broken down by category. For the project-wide value alone, use `scope: categories` and read the entry whose `category_id` is null. Use this for the trend question ("is the project gaining or losing visibility?"). Use the get_project_visibility_by_domain tool when you need the value per individual domain. `scope` controls the level of detail: `categories` (default) adds the per-category breakdown to every date, `project` returns only the total per date and is much smaller — prefer it when you just need the curve. A `scope` other than these two strings, a `months` that is not an integer of at least 1 (values above 48 are capped) and a `category_id` that is not a category of this project a…
-
generate_content_textGenerate Content Text**Consumes 1 content AI credit per call** (the Content Suite budget `content_suite.ai_generation`, see the get_account_credits tool). Runs one of the Content Suite writing assistants and returns its answer as plain text — the API equivalent of the AI helpers in the Quaro editor. `type` selects the assistant, `message` is the instruction you would type in the app. Everything the assistant needs (keyword, language, target audience, existing text) must be part of `message`; there are no separate fields for it. Use the generate_article_from_outline tool instead when you want a complete, multi-section article from an outline. The credit is reserved when the call starts, so parallel calls cannot exceed the limit. It is given back only when the generation could not be started at all; a generation that was started but did not finish in time keeps running in the background and stays charged. This request can run up to 120 seconds. With a personal API key this endpoint needs the Content Suite f…
-
analyze_tone_of_voiceAnalyze Tone Of Voice**Consumes 1 content AI credit per call** (the Content Suite budget `content_suite.ai_generation`, see the get_account_credits tool). Derives the tone of voice from a sample text and returns it as a short description. Typical workflow: send an existing text of the brand here, then pass the returned `tone_of_voice` into the generate_article_from_outline tool so the generated article matches the brand voice. The credit is reserved when the call starts and given back when the analysis fails (502). Stateless — nothing is stored on a project. This request can run up to 60 seconds. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted.
-
generate_article_from_outlineGenerate Article From Outline**Consumes content AI credits** (the Content Suite budget `content_suite.ai_generation`, see the get_account_credits tool) — 1 credit per section of the project's outline, but at least 1 per 1,500 words of the project's target word count, i.e. `max(number of headings, ceil(target word count / 1500))`. An outline of 8 headings with a target of 1,500 words costs 8 credits; 2 headings with a target of 20000 words cost 14. The credits are reserved when the call starts, so parallel calls cannot exceed the limit; what is not used is given back. **Partial result:** if a section fails after at least one was written, the call still answers 200 with the finished sections plus `complete: false`, `sections_requested` and `missing_headlines`. The finished sections stay charged, and so does the section the run failed on when its generation had already been started (it keeps running in the background); only sections that were never started are given back (`meta.credits_consumed` / `meta.credits_rele…
-
rename_research_collectionRename Research Collection**Write operation.** Renames a collection. The keywords, categories and custom columns are kept; only the name changes. All other collection endpoints address the collection by name, so follow-up calls must use the new name. The new name must be unique for the account. The rename is confirmed by reading the collection back under its new name before answering; a 502 means the rename was started but could not be confirmed — check the list_research_collections tool before retrying. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additional_info` says how many seconds ago the import last reported progress. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 2 (rename); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the…
-
create_collection_columnCreate Collection Column**Write operation.** Defines a typed custom column on a collection — the API equivalent of the ADD COLUMN dialog of the collection table. Every existing keyword is seeded with the type's default value (dropdown: the "-" option, text and number: empty), which is what makes the column render in app.quaro.io. Three types are available: `text` (free text), `number` (numeric values) and `dropdown` (a fixed set of options, each optionally with a color). For a dropdown, the "-" option is added automatically as the unset state — do not include it yourself. After creation, write values with the set_collection_column_value tool, inspect the defined columns with the list_collection_columns tool, and remove a column with the delete_collection_column tool. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additio…
-
list_collection_columnsList Collection ColumnsLists the custom columns defined on a collection with their type, description and — for dropdown columns — the selectable option texts. Use this to learn what the set_collection_column_value tool will accept. Does not consume credits. Personal keys need the Research Suite right id 1 (see collections and research projects); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
delete_collection_columnDelete Collection Column**Write operation.** Deletes a custom column defined on a collection (see the create_collection_column tool), removing both its header definition AND the value it holds on every keyword in the collection — not just the column definition. This cannot be undone. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additional_info` says how many seconds ago the import last reported progress. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 40 (delete columns); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
list_research_projectsList Research ProjectsLists the research projects of the account. A research project is the folder level above collections: it groups collections, e.g. per customer or per topic, without changing the collections themselves. Use the returned `name` as the `project` attribute of the other research project endpoints. Does not consume credits. Personal keys need the Research Suite right id 1 (see collections and research projects); without it the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
create_research_projectCreate Research Project**Write operation.** Creates a research project — a folder that groups collections. At least one existing collection is assigned right away (like in the app); more can be added later with the assign_collection_to_research_project tool. The project name must be unique for the account. Does not consume credits. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 30 (create research projects); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted. Takes NO keywords — only the NAMES of collections that already exist (via assign_collection_to_research_project or the collections param). To create new keywords with search volume, use create_research_collection instead; group it into a project afterwards.
-
assign_collection_to_research_projectAssign Collection To Research Project**Write operation.** Assigns an existing collection to a research project. A collection belongs to at most one project, so — like "Add to project" in app.quaro.io — this is a MOVE: a collection that already sits in another project is taken out of that project first (`moved_from` names it). The collection itself is not changed — it only changes its project grouping, and the keyword counts of every project involved are updated. The assignment (and the removal from the previous project) is verified before the response; a 502 means it could not be confirmed. Does not consume credits. Personal keys need the Research Suite right id 1 (see collections and research projects) plus right id 30 (create research projects) or right id 34 (add collections to research projects); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
remove_collection_from_research_projectRemove Collection From Research Project**Write operation.** Detaches a collection from a research project. The collection itself survives untouched — only the assignment is removed and the project's keyword count is corrected. To delete a collection permanently, use the delete_research_collection tool instead. Does not consume credits. Personal keys need the Research Suite right id 1 (see collections and research projects) plus either right id 3 (delete any collection) or right id 28 (delete own collections, only as the owner of the collection); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
delete_research_projectDelete Research Project**Write operation.** Deletes a research project. Only the grouping is removed — the collections it contained are NOT touched and remain available. To delete one of those collections permanently, use the delete_research_collection tool instead. This cannot be undone. Does not consume credits. While the project is being created or updated the request is rejected with a 400 on `project`. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 31 (delete research projects); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
create_monitoring_projectCreate Monitoring Project**Write operation.** Creates a Google ranking (monitoring) project and imports its keywords in one call — the API equivalent of the monitoring setup wizard in app.quaro.io. **Consumes Monitoring Suite keyword credits:** unique keywords × updates per month (interval 1 → 30, 7 → 4, 14 → 2, 28 → 1), doubled when `ai_overview` is true. Example: 100 keywords, weekly, `ai_overview` false = 400. The response echoes the reservation as `reserved_keyword_updates`. The project is stored exactly as the app's setup wizard stores it: keywords are lowercased and special characters replaced by a space (max. 80 characters / 10 words each), `serp_depth` is fixed at 20, and for the app's own locations (e.g. 2756 Switzerland / de) the app's location label is stored. When the account has too few credits, the endpoint answers 400 with status_code 50000 and nothing is created. The response arrives as soon as the project exists — the keyword import continues in the background (`import_status: "running"`) and…
-
update_monitoring_projectUpdate Monitoring Project**Write operation.** Changes the settings of an existing monitoring project. Only the sent fields are changed; at least one updatable field is required. Updatable: `title`, `domain`, `interval`, `categorized_project`, `ai_overview` — the same fields the app's project-settings dialog edits. `ai_overview` (JSON boolean) switches loading the content of AI Overviews on or off. On doubles the reserved keyword updates; the credit difference is checked and reserved like an interval change (same rejection when the credits do not cover it, same refusal during an import). The content is loaded from the next update on; AI Overviews already stored stay without content. Sending the stored value is a no-op. `interval` (1 = daily, 7 = weekly, 14 = every two weeks, 28 = every four weeks) changes the keyword credits reserved for the project. The difference is checked and reserved before anything is saved; if the credits do not cover it, the whole update is rejected with 400 / 50000 'Out of credits.' a…
-
delete_monitoring_projectDelete Monitoring Project**Write operation.** Permanently deletes a monitoring project together with all its keywords, rankings, competitors, categories and clusters. This cannot be undone. The keyword credits the project reserved are released at the end of the already-paid period, not immediately — see `credits_release`/`hint` on the response and `pending_release` on the get_account_credits tool (`monitoring_suite.pending_release`). Only works on monitoring projects — deleting an AI Suite project goes through the delete_ai_project tool. While keywords are still being imported or an interval change is being applied, the delete is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`) and nothing is deleted; retry later. This request can run up to 180 seconds. A demo project in the account (an example project that is not updated and does not count towards the credits) can be read but not changed; a write is refused with 400 on `project_id` (status_code 55017, `additional_info` na…
-
create_ai_projectCreate Ai Project**Write operation.** Creates an AI Suite project that tracks how visible a brand is in AI answers, and imports its prompts in one call — the API equivalent of the AI Suite setup wizard. **Consumes AI Suite prompt credits:** prompts × updates per month (interval 1 → 30, 7 → 4, 14 → 2, 28 → 1) × number of platforms. The response echoes the reservation as `reserved_prompt_updates`. When the account has too few credits, the endpoint answers 400 with status_code 50000 and nothing is created. Each entry in `prompts` may be a plain string, or a `{prompt, tags}` object to tag it on import (see the set_ai_prompt_tags tool to retag already-tracked prompts later) — the two forms can be mixed in one array. Tagging at least one prompt this way also sets the project's `categorized_project` flag, mirroring the AI Suite setup wizard. A call where no entry carries tags behaves exactly like the plain-string form. The response arrives as soon as the project exists — the prompt import continues in the ba…
-
delete_ai_projectDelete Ai Project**Write operation.** Permanently deletes an AI Suite project together with its prompts, stored AI answers and brand data. This cannot be undone. Calling this endpoint does not consume credits. The prompt credits the project reserved are released at the end of the already-paid period, not immediately — see `credits_release`/`hint` on the response and `pending_release` on the get_account_credits tool (`ki_suite.pending_release`). Only works on AI Suite projects — deleting a monitoring project goes through the delete_monitoring_project tool. While keywords, prompts or a platform are still being imported, or an interval change is being applied, the delete is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`) and nothing is deleted; retry later. This request can run up to 180 seconds. A demo project in the account (an example project that is not updated and does not count towards the credits) can be read but not changed; a write is refused with 400 on `pr…
-
add_ai_promptsAdd Ai Prompts**Write operation.** Adds prompts to an existing AI Suite project. Deduplicates twice: within the submitted `prompts` list, AND against prompts already tracked on the project (exact text match after trimming) — a prompt that already exists is skipped, not re-added or re-charged. **Consumes AI Suite prompt credits** only for the prompts actually added (prompts × updates per month × platforms, where platforms includes AI Overview if enabled). The response echoes the reservation as `reserved_prompt_updates`. When the account has too few credits, the endpoint answers 400 with status_code 50000 and nothing is added. Each entry in `prompts` may be a plain string, or a `{prompt, tags}` object to tag it on import (see the set_ai_prompt_tags tool to retag already-tracked prompts later) — the two forms can be mixed in one array. A call where no entry carries tags behaves exactly like the plain-string form. The new prompts are answered with the project's next scheduled update, not immediately (`…
-
set_ai_prompt_tagsSet Ai Prompt Tags**Write operation.** Sets, adds, or removes tags on already-tracked prompts of an AI Suite project, identified by their exact prompt text (see the get_ai_prompts tool for the tracked prompts). Existing response history for these prompts is untouched — this is the way to fix or add tags after import (e.g. a CSV import that came in without tags) without deleting and re-adding the prompts, which would lose their accumulated answers. Provide exactly one of `set_tags` (replaces a prompt's whole tag list — `set_tags: []` clears all tags), `add_tags`, or `remove_tags`. Every tag is trimmed; empty and duplicate tags are dropped. A tag longer than 200 characters is rejected (400 on the field carrying it). `add_tags`/`remove_tags` must contain at least one non-empty tag; only `set_tags` may be empty. Does not consume credits. A demo project in the account (an example project that is not updated and does not count towards the credits) can be read but not changed; a write is refused with 400 on `…
-
delete_ai_promptsDelete Ai Prompts**Write operation.** Deletes prompts from an AI Suite project, identified by their exact prompt text (see the get_ai_prompts tool for the tracked prompts). The stored answers of these prompts are removed with them and the credits they reserved are released at the end of the already-paid period, not immediately — see `credits_release`/`hint` on the response and `pending_release` on the get_account_credits tool (`ki_suite.pending_release`). This cannot be undone. Calling this endpoint does not consume credits. `deleted` reports the number ACTUALLY removed — a text that was never tracked (or already deleted) does not count toward it. Compare against `requested` to see the difference. A demo project in the account (an example project that is not updated and does not count towards the credits) can be read but not changed; a write is refused with 400 on `project_id` (status_code 55017, `additional_info` names the reason).
-
create_content_projectCreate Content Project**Write operation.** **Consumes one Content Suite project credit** (`content_suite.new_projects` in the get_account_credits tool). Creates a content project — the container for briefing, editor and WDF*IDF work on one piece of content, addressed by a main keyword. The project is created **empty**: topics, competitors and WDF*IDF data are produced by the setup_content_project tool (or the step-wise the setup_content_project_topics tool → /setup/competitors → /setup/run), not by this call. Use the run_wdfidf_analysis tool for an ad-hoc analysis without a project. The title must be unique among the account's content projects. Location and language have to be one of the nine pairs the app offers for a content project — see `location_code`. The credit is reserved when the call starts and given back when no project is created (for example because the title is already taken), so parallel calls cannot exceed the limit. The new project also counts towards `content_suite.active_projects` until…
-
update_ai_projectUpdate Ai Project**Write operation.** Changes the settings of an existing AI Suite project. Only the sent fields are changed; at least one updatable field is required. Updatable: `title`, `brand_name`, `brand_domain`, `interval`. `interval` (1 = daily, 7 = weekly, 14 = every two weeks, 28 = every four weeks) changes the prompt credits reserved for the project. The difference is checked and reserved before anything is saved; if the credits do not cover it, the whole update is rejected with 400 / 50000 'Out of credits.' and nothing is changed. While an import is running the change is refused with 400 on `project_id` (status_code 55017, the reason in `additional_info`); retry later. On a paused project the new interval is saved and its credits are reserved when the project is resumed. Sending the stored interval (or 28 on a legacy 30/31-day project) is a no-op. When the interval changed, the response also contains `interval`, `previous_interval`, `paused`, `monthly_prompt_updates` {before, after} (not fo…
-
add_ai_project_platformsAdd Ai Project Platforms**Write operation.** Adds AI platforms to an existing AI Suite project. **Consumes AI Suite prompt credits** for every prompt already in the project on every newly added platform. The reservation is prompt_count × updates per month (1→30, 7→4, 14→2, 28→1) × number of NEWLY added platforms. The response echoes the actual figure as `reserved_prompt_updates` (0 while the project is paused). Platforms already tracked on the project are ignored; if every given platform is already tracked, the request fails with a validation error instead of doing nothing silently. The project's initial import must have finished first — while it is still running the endpoint answers 400 with a hint to retry later. The same applies while prompts are being imported, another platform is being added, an interval change is being applied, or an update or answer fetch is running: 400 on `project_id` (status_code 55017, the reason in `additional_info`), nothing is reserved; retry later. The new platform is answered…
-
update_research_projectUpdate Research Project**Write operation.** Changes the settings of an existing research project. Only the sent fields are changed; at least one updatable field is required. Updatable: `new_name`, `description`, `customer`. The new name must be unique among the account's research projects. Does not consume credits. While the project is being created or updated the request is rejected with a 400 on `project`. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 32 (edit research projects); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
update_content_projectUpdate Content Project**Write operation.** Changes the settings of an existing content project. Only the sent fields are changed; at least one updatable field is required. Updatable: `url`, `briefing_description`, `word_count_target`, `term_usage`, `img_count`, `readability`. `device` is fixed at creation and can only be sent back unchanged (see the field). `word_count_target`, `term_usage`, `img_count` and `readability` are the targets the editor's content score grades against — changing them does not rewrite existing content, only what "good" means for it going forward. All four can be sent either flat (`word_count_target`, `term_usage`, `img_count`, `readability` — the shape the list_content_projects tool reports for the first two) or nested under `settings` (`settings.word_count`, `settings.term_usage`, `settings.img_count`, `settings.readability` — the shape the full the get_content_project tool returns unchanged): either GET response can be fed straight back into this endpoint without reshaping it. I…
-
setup_content_projectSetup Content Project**Write operation.** Runs the setup pipeline that turns an empty project from the create_content_project tool into a usable WDF*IDF project — the API equivalent of clicking through the setup wizard steps in app.quaro.io. Until the setup has finished, the project is not shown in the app. Runs, in order: (1) derive a topic cluster from the main keyword via a live SERP call, (2) select topics, (3) find competitors per selected topic via a live SERP call, (4) select competitors, (5) scrape the selected competitors and compute WDF*IDF. Steps 2 and 4 are not optional — the analysis only covers what gets selected, so skipping them would analyse an empty set. By default ONE topic is selected - the one that equals the project's main keyword (else the first discovered topic), and that topic's keyword becomes the project's main keyword, exactly like the app's single-select topic table; send `max_topics` (1-5) to take more topics in discovery order. Whenever exactly one topic ends up selected — a…
-
setup_content_project_topicsSetup Content Project Topics**Write operation.** Step-wise counterpart of the setup_content_project tool, mirroring the topic step of the setup wizard in app.quaro.io. Runs a live SERP analysis for the project's `main_keyword` and derives a topic cluster — nothing is selected yet, that happens in the next step, the setup_content_project_competitors tool. The topic that equals the main keyword is listed first and named as `recommended_topic` (see also `meta.note`), because selecting a single topic in the next step makes its keyword the project's main keyword. Use the step-wise routes instead of the one-shot the setup_content_project tool when you want to inspect or curate topics/competitors between steps; both paths use the same selection logic and produce an identical end result. You can freely mix "one call" and "three calls" on the same project, and call the steps in separate requests (the progress is stored on the project). **Topics are reused:** if the project already has topics, this call returns them as th…
-
setup_content_project_competitorsSetup Content Project Competitors**Write operation.** Step-wise counterpart of the setup_content_project tool, mirroring the competitor step of the setup wizard in app.quaro.io. Requires the setup_content_project_topics tool to have run first. Selects topics from the given `topics` list, then runs one live SERP fetch per selected topic to discover competitors and PAA-style questions. **Selecting exactly one topic makes its keyword the project's main keyword** (like the app's single-select topic table) — the response reports the resulting `main_keyword` and, if it changed, `main_keyword_changed_from`; with several topics the main keyword stays as it is. The discovery preselects the top `depth` competitors (chosen at the create_content_project tool) by average SERP position, returned as `selected`; the final selection happens in the next step, the run_content_project_setup tool, which keeps that preselection when called without `competitors`/`max_competitors`. **`topics` is required.** The app's own wizard starts this…
-
run_content_project_setupRun Content Project Setup**Write operation.** Step-wise counterpart of the setup_content_project tool, mirroring the final analysis step of the setup wizard in app.quaro.io. Requires the setup_content_project_competitors tool to have run first. Selects competitors (explicit `competitors`, the first `max_competitors` in discovery order, or — when both are omitted — the discovery's preselection: top `depth` by average SERP position), scrapes each selected competitor page and computes the WDF*IDF keyword analysis — the same completion this route's one-shot sibling the setup_content_project tool performs. Marks the project finished: only after this step does the project appear in app.quaro.io and become usable for editor and briefing calls. **`fastmode` projects cannot call this route at all:** the setup_content_project_competitors tool already finishes a `fastmode:true` project (scrapes the auto-selected competitors, runs WDF*IDF, marks it `finished` — see that route's `setup_complete` response field). By the ti…
-
setup_content_project_termsSetup Content Project Terms**Write operation.** Sets the SELECTION SET of terms after the run_content_project_setup tool has analyzed the project — the API equivalent of the term-selection step in app.quaro.io's setup wizard. This is a full replacement of the selection: every term the analysis produced ends up either selected or not, matching exactly the `terms` list sent (unlike the setup_content_project_questions tool' `add_questions`, there is no way to add a brand-new term here — terms only come from the analysis). The analysis itself preselects the 20 terms with the highest average WDF*IDF; that selection stays until it is changed here or in the app. Identity is the exact term text (case-insensitive), matched against the terms the run_content_project_setup tool already stored on the project (see the get_content_project_keywords tool for the full analyzed list). Does not consume credits. With a personal API key, a project whose setup is not finished can only be worked on by the user who created it or by an…
-
setup_content_project_questionsSetup Content Project Questions**Write operation.** Sets the SELECTION SET of discovered questions and/or adds the caller's own custom questions to a content project's briefing — the API equivalent of the question-selection step in app.quaro.io's setup wizard. Discovered questions come from two producers (SERP "people also ask" results and competitor content extraction), both stored with `selected:false` until this endpoint (or the app) picks some. Provide at least one of `questions`/`add_questions`. `questions` is a full replacement of the selection among already-discovered questions (every discovered question not named becomes unselected) — mirrors the setup_content_project_terms tool. `add_questions` inserts brand-new questions of your own, always selected and flagged `user_input:true`, deduplicated case-insensitively after trimming against what's already there (and against each other); both can be sent in the same call. Questions are identified by their exact text (case-insensitive) — they have no ids. With a p…
-
delete_content_projectDelete Content Project**Write operation.** Permanently deletes a content project together with its keywords, editor document, editor history and workflow, and removes it from its folders. This cannot be undone. Lowers the account's active-project count (`content_suite.active_projects` in the get_account_credits tool) by one, reported as `meta.credits_released`; the project credit charged at the create_content_project tool is not given back. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted. With a personal API key, a project whose setup is not finished can only be worked on by the user who created it or by an admin; for anyone else the call answers 400 as for an unknown `project_id`. Before calling, read the project's title back to the user (from list_content_…
-
create_ai_project_reportCreate Ai Project Report**Write operation.** Creates a saved report for an AI Suite project — a curated, brandable slice of the project's dashboard that can later be published behind a public link. A report starts with the same defaults as "New report" in the app: template "brand_visibility"; without `sections` it gets the 4 default sections (kpi_cards, visibility_trend, top_brands_per_platform, brand_table, all on platform "all"); without `filters.selectedDates` it gets the app's 30-day preset (the first processed day within the last 30 days up to the last processed day of the project). What you send is checked against the project the way the app's report builder restricts it: a section's `platform` must be "all" or a platform the project tracks, a section's `tag` must occur in the project's AI responses, and `filters.selectedDates` must contain at least one tracked day. Anything you send wins over these defaults — send `sections`/`filters`/`branding`/`summary`/`language` right away or change them later wit…
-
update_ai_project_reportUpdate Ai Project Report**Write operation.** Patches a report. Only the fields you send are changed — every omitted field keeps its current stored value. At least one updatable field is required: `name`, `template`, `sections`, `filters`, `branding`, `summary`, `language`. Note that `sections` is replaced as a WHOLE array when sent, not merged item-by-item — send the full desired list of sections, not just the one you want to add or change. Read the report first with the get_ai_project_report tool if you need to keep some existing sections while changing others. `filters` is also replaced as a whole object, with one exception: a `filters` object without `selectedDates` keeps the stored date range. `branding` is merged key by key: sending only `primaryColor` keeps the stored `logoUrl` (send `logoUrl: null` to remove the logo). What you send is checked against the project exactly like the app's report builder restricts it: a section's `platform` must be "all" or a platform the project tracks, a section's `tag`…
-
list_ai_project_reportsList Ai Project ReportsLists the reports saved on an AI Suite project. Use the returned `report_id` with the other report endpoints. Does not consume credits.
-
get_ai_project_reportGet Ai Project ReportReturns one report in full, including all its sections, filters and branding. Does not consume credits.
-
delete_ai_project_reportDelete Ai Project Report**Write operation.** Permanently deletes a report. If its public share link was enabled, the link stops working immediately. This cannot be undone. Does not consume or release credits. For a demo project in the account the reports can be read but not created, changed, deleted or published (400 on `project_id`, status_code 55017). Permanent. If the report's public link was enabled, it stops working immediately. Confirm the report name with the user first.
-
publish_ai_project_reportPublish Ai Project Report**Write operation.** Publishes a report behind a public, unauthenticated link that anyone with the URL can view — no Quaro account or API key required on their end. Enabling an already-shared report mints a FRESH link and invalidates any link that was handed out before. There is no way to reuse an old link once you call this again. Does not consume credits. For a demo project in the account the reports can be read but not created, changed, deleted or published (400 on `project_id`, status_code 55017). Exposes the report behind a PUBLIC, unauthenticated link — anyone with the URL can view it, no Quaro account needed. Confirm with the user before calling. Calling it again on an already-shared report mints a FRESH link and invalidates any link handed out before (unlike the content/collection publish tools, which are idempotent and keep the existing link).
-
unpublish_ai_project_reportUnpublish Ai Project Report**Write operation.** Disables the public link of a report. The report and its data are untouched — only the public link stops resolving. Publish again with the publish_ai_project_report tool to get a new link. Does not consume credits. For a demo project in the account the reports can be read but not created, changed, deleted or published (400 on `project_id`, status_code 55017).
-
get_ai_opportunitiesGet Ai OpportunitiesThe Opportunities tab of an AI project: tracked prompts whose AI answers leave your brand out while competitors are named, ranked by an opportunity score (0-100: share of answers without you, competitor pressure, missing in the latest run, number of platforms). Each row has its gap type (`never` named, `lost` in the latest run, `platform_gap` named on some platforms only, `partial` named in part of the answers), the competitors named instead, the domains cited, and its worklist `status` and `note`. Besides the rows the answer carries the KPIs of the window and the previous one, the trend (new / resolved / persisting gaps), counts per status and gap type, topics (tags) with your presence share, and the competitors that win the most of your gaps. The first request for a project (and the first after a new import) may answer `preparing: true` with a `progress` while the index is built in the background — ask again after `retry_after_seconds`. By default only "strong" gaps are listed (neve…
-
get_ai_opportunity_detailGet Ai Opportunity DetailFor one prompt of the get_ai_opportunities tool: the brands named and the sources (domains, URLs, page types) and search queries the models used in exactly the answers where your brand is missing. Same window and filters as the list. A project whose first import is still running is refused with 400 on `project_id` (status_code 55017, `additional_info` says that the project is still importing), like on the other data endpoints of an AI project. Does not consume credits.
-
get_ai_opportunity_action_planGet Ai Opportunity Action PlanTurns the gaps of a project into tasks, the "Action plan" of the Opportunities tab: write or improve an own page for a group of prompts (with example pages the models cite today), get mentioned on a page that is cited for several of your gap prompts, or answer a negative statement. Each task has a `key`, a priority, the prompts it covers and its worklist status; set the status with the set_ai_opportunity_status tool and the task `key` as `prompt`. Without `tag` the plan covers all topics; with `tag` only that topic. May answer `preparing: true` like the list. A project whose first import is still running is refused with 400 on `project_id` (status_code 55017, `additional_info` says that the project is still importing), like on the other data endpoints of an AI project. Does not consume credits. If the answer says preparing: true, wait retry_after_seconds and call again.
-
set_ai_opportunity_statusSet Ai Opportunity Status**Write operation.** Sets the worklist status (and note) of a prompt from the get_ai_opportunities tool or of a task from the action plan (pass the task `key` as `prompt`). The status is shared by everyone on the project, exactly like in the app. Not allowed for the read-only AI Suite role. `prompt` has to be exactly the prompt text of a row of the get_ai_opportunities tool or the `key` of an action plan task. It is looked up over the 365 days up to the newest run of the project, so a row from an earlier window than the newest 92 days (the widest single `date_scope` of the opportunities) can be set as well, like in the app's "Last 12 months" view; anything else is rejected with 400 55029 on `prompt`. While the opportunities of the project are still being prepared the lookup is not possible yet and the request is answered with 400 on `prompt` as well - retry after a few seconds. An action plan task can only be set to `open` or `ignored` (400 55029 on `status` otherwise), as in the app.…
-
list_ai_page_auditsList Ai Page AuditsLists the Page Audit runs of the account, newest first (optionally only those of one AI project), together with the audit credits used and available this month. A Page Audit checks one page for how well AI answers can find, read and quote it — not to be confused with the audit_page tool, a technical on-page crawl. Does not consume credits.
-
get_ai_page_auditGet Ai Page AuditOne Page Audit run and its result: the total and partial scores (technical, structure, machine readability, citability), every check with pass/warn/fail and its impact, the recommendations, quotable passages and E-E-A-T signals, the coverage of the project prompts / focus topics / fan-out queries (covered, partial, missing with the passage or the gap), how often the page is cited in the project's AI answers, and the comparison with the previous audit of the same URL. A run takes about one to three minutes. While it runs (`status` pending or processing) the answer has `retry_after_seconds` and no results yet. Does not consume credits. While status is pending or processing there are no results yet — wait retry_after_seconds before asking again, at most a few times.
-
create_ai_page_auditCreate Ai Page Audit**Write operation, consumes 1 Page Audit credit** - a separate monthly AI Suite allowance (not the prompt credits; the get_account_credits tool does not show it). `credits` on the list_ai_page_audits tool reports used, limit and available. The credit is booked only when the page was audited successfully; with no credit left the endpoint answers 400 with status_code 50000 and no audit starts. Starts a Page Audit of ONE page and returns its `run_id`; fetch the result with the get_ai_page_audit tool after `retry_after_seconds`. With `project_id` the page is also checked against the prompts tracked in that AI project. `focus` adds up to 20 keywords or prompts the page should be found for. Not allowed for the read-only AI Suite role. With the `project_id` of a demo project in the account the audit is refused (400 on `project_id`, status_code 55017) and no credit is used. Costs 1 audit credit (see list_ai_page_audits for the credits left) — only on the user's request. The audit runs in the…
-
delete_ai_page_auditDelete Ai Page Audit**Write operation.** Permanently deletes a Page Audit run and its results. Calling this endpoint does not consume credits; audit credits already used are not refunded. This cannot be undone. Not allowed for the read-only AI Suite role. An audit that belongs to a demo project in the account cannot be deleted: demo projects are read-only (400 on `project_id`, status_code 55017). Deletes one page audit - only on the user's request. An audit that belongs to a demo project cannot be deleted (demo projects are read-only).
-
save_content_project_editorSave Content Project Editor**Write operation.** Writes the editor content and/or the SERP snippet (title/description) of a content project — the API equivalent of typing in the app.quaro.io content editor and its snippet fields. **Blocked while a human has the editor open:** check `locked` on the get_content_project_editor tool first. If someone was active in the editor within the last 15 seconds, this call answers 400 instead of overwriting what they are typing — retry a little later. Your own save through this endpoint never blocks your next call. Content and snippet fields save independently — send only what changed. `content`/`comments` are saved together as the editor document; `title`/`description` are saved together as the SERP snippet. The stored `content_score` (readability, keyword coverage, image count) is not changed by this endpoint. **The content score is only (re)calculated by the Quaro editor in the browser:** after an API save, the score ring in the project list keeps its previous value until s…
-
set_content_workflow_stageSet Content Workflow Stage**Write operation.** Advances a content project's ACTIVE stage to one of its workflow's already-defined stages — the API equivalent of dragging the project to another column on the editorial workflow board. This only changes which stage is "current"; it does not complete any tasks (use the complete_content_workflow_task tool for that). Not sticky, like in app.quaro.io: the next workflow write (workflow/update, stages/add|remove, tasks/add|remove|assign|complete|comment) recomputes the active stage from task completion — the first stage with an open task, or "Finished" — and overwrites this choice. The project must already have a workflow — create one first with the create_content_workflow tool if it does not. Address the target stage by its title, matched case-insensitively (see the get_content_project_workflow tool for the available stage titles). Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `…
-
publish_content_project_sectionPublish Content Project Section**Write operation.** Publishes one section of a content project behind a public, unauthenticated guest link that anyone with the URL can view — no Quaro account required on their end. Mirrors the "Share" toggle in app.quaro.io for that section. Idempotent: calling it again on a section that is already shared re-confirms the same state and answers with `shared: true`, it does not rotate or invalidate the existing link (unlike the publish_ai_project_report tool, which always mints a fresh token). Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted. With a personal API key, a project whose setup is not finished can only be worked on by the user who created it or by an admin; for anyone else the call answers 400 as for an unknown `project_id`.…
-
unpublish_content_project_sectionUnpublish Content Project Section**Write operation.** Disables the public guest link of one section of a content project. The project and its data are untouched — only the public link stops resolving. Idempotent: calling it on a section that is already unshared answers with `shared: false` and changes nothing. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted. With a personal API key, a project whose setup is not finished can only be worked on by the user who created it or by an admin; for anyone else the call answers 400 as for an unknown `project_id`.
-
publish_research_collectionPublish Research Collection**Write operation.** Publishes a collection behind a public, unauthenticated guest link that anyone with the URL can view — no Quaro account required on their end. Mirrors the "Share" toggle in app.quaro.io. Idempotent: calling it again on a collection whose public link is already active re-confirms the same state and answers with `shared: true`. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additional_info` says how many seconds ago the import last reported progress. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 4 (share); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted. Exposes the collection behind a PUBLIC, unauthenticated link —…
-
unpublish_research_collectionUnpublish Research Collection**Write operation.** Disables the public guest link of a collection. The collection and its keywords are untouched — only the public link stops resolving. Idempotent: calling it on a collection whose public link is already off answers with `shared: false` and changes nothing. Does not consume credits. While an import is running on the collection (status `creating` or `updating` in the list_research_collections tool), this request is rejected with a 400 (55029) on `collection`, like the app hides this action; `additional_info` says how many seconds ago the import last reported progress. Personal keys need the Research Suite right ids 1 (see collections and research projects) and 4 (share); without them the call is refused with 403 / status_code 55024 — see `research_rights` in the get_account_me tool. Admins and the company-wide key are not restricted.
-
create_content_workflowCreate Content Workflow**Write operation.** Creates the editorial workflow of a content project — named stages, each with named tasks that can be assigned, due-dated, completed and commented on. A project can have at most one workflow; use the update_content_workflow tool to change an existing one. Omit `stages` to seed the app's own default template (Briefing → Production → Lectorate → Publishing, with two tasks each) — the result is indistinguishable from a workflow created in app.quaro.io. Send `stages` to define a custom structure instead. Stages and tasks are addressed everywhere else in the API by their **title**, not by an internal id — titles must therefore be unique within their scope (stage titles within the workflow, task titles within their stage). Assignees are addressed by **email**; every email must belong to a Content Suite user of this account (see the list_content_suite_users tool) or the request fails. Omit `title` for the default title "Default Content Workflow"; a sent `title` has to be…
-
update_content_workflowUpdate Content Workflow**Write operation.** Changes the title and/or the stage/task structure of an existing workflow. At least one of `title` or `stages` is required. The project must already have a workflow — create one first with the create_content_workflow tool. Sending `stages` replaces the ENTIRE structure — always send the full stage/task list you want, not just what changed. Despite being a full replacement, **progress is preserved**: for every task whose (stage title, task title) still exists after the update, its completion state, completion date and comments are carried over automatically, its assignees are kept unless the new task sends a non-empty `assignees` list (an empty `assignees: []` also keeps the stored ones — clear them with the assign_content_workflow_task tool), and its due date is kept unless the new task sends `due_date` (`null` clears it). Renaming a stage or task title is the one case where this cannot apply — a rename is indistinguishable from removing the old task and adding a…
-
assign_content_workflow_taskAssign Content Workflow Task**Write operation.** Sets who is responsible for one workflow task. `assignees` is the COMPLETE list for that task, not an addition — send every email that should remain assigned, including the ones already there; an empty array clears the task's assignees entirely. Every email must belong to a Content Suite user of this account (see the list_content_suite_users tool) or the request fails naming the offending address. Like in the app, the task's assignees are notified by email and are added to the project's authors — that is what makes the project visible to a user with the Author role. `stage`, `task` and `assignees` are all required — a missing `stage` or `task` is rejected with its own message, and `assignees: null` is rejected like an omitted list (send `[]` to clear). Unlike the update_content_workflow tool, an empty array here DOES clear the task's assignees. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects to…
-
complete_content_workflow_taskComplete Content Workflow Task**Write operation.** Marks one workflow task as completed or reopens it. Completing stamps the completion date the same way app.quaro.io does; reopening clears it again. Like every workflow write, it recomputes the project's active stage from task completion (as the app does) and reports it as `current_stage`. When a task is newly completed, everyone involved in the workflow (the assignees of its tasks and everyone who commented) is notified by email, like in the app — except the user of the personal key that made the call. Reopening a task sends nothing; completing a task that is already completed sends nothing either and keeps its original completion date. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects tool); otherwise the call answers 400 as for an unknown `project_id`. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access or Author role (`suites.content_suite.role`…
-
comment_on_content_workflow_taskComment On Content Workflow Task**Write operation.** Adds a comment to a workflow task. Like in the app, everyone involved in the workflow (the assignees of its tasks and everyone who commented before) is notified by email with the comment text. With a personal API/MCP key (per-user, not the legacy account-wide key), the comment author is always the key's own owner — omit `author` entirely; supplying a different email is rejected with a 400 (status_code 55029, attribute `author`). The legacy account-wide key identifies the account, not a person — a comment made with it is stored with the neutral author `API` (reported as `api` in the `author` of the comment), `author` is ignored and no longer needed, and the notification to the people involved names `API` instead of a colleague. The response then carries a `meta.note` saying so; use a personal key to comment in your own name. With a personal API key the project has to be visible to the user of that key (same rules as the list_content_projects tool); otherwise the ca…
-
list_content_suite_usersList Content Suite UsersLists the email addresses of this account's Content Suite users — the only people the API can name at all. Use these emails as `assignees` in the create_content_workflow tool, the update_content_workflow tool and the assign_content_workflow_task tool. Only a user's email is returned — no internal id, permission flags or display name. Does not consume credits.
-
delete_content_workflowDelete Content Workflow**Write operation.** Permanently removes a content project's workflow — its stages, tasks, assignments, completion state and comments. The content project itself and its content are untouched, and a new workflow can be created afterwards with the create_content_workflow tool. This cannot be undone. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted. With a personal API key, a project whose setup is not finished can only be worked on by the user who created it or by an admin; for anyone else the call answers 400 as for an unknown `project_id`. Permanently removes the workflow's stages, tasks, assignments and comments (the content itself is untouched). Confirm with the user first.
-
add_content_workflow_stageAdd Content Workflow Stage**Write operation.** Adds ONE new stage to an existing workflow without touching any existing stage — the safe alternative to the update_content_workflow tool when all you want is to insert a stage: because nothing else is sent, nothing else can be dropped by omission. The stage's title can be sent as `stage` or `title` (see the schema). The stage must not already exist; to add tasks to an existing stage, use the add_content_workflow_task tool instead. A workflow has at most 20 stages — adding to a full one is rejected with a 400 on `stage` ("The workflow already has 20 stages, the maximum — remove one first."). Errors about the stage title come back with `attribute: "stage"`; everything about its tasks (count, titles, assignees, due dates) with `attribute: "tasks"` and the location (e.g. `tasks[0].due_date`) in `additional_info`. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_a…
-
remove_content_workflow_stageRemove Content Workflow Stage**Write operation.** Removes ONE stage from a workflow, together with every task in it — its progress, assignments and comments go with it. This is the one endpoint in the stage/task membership group that is destructive by nature, since a stage IS its tasks; the remove_content_workflow_task tool only ever removes the one task named. Refused on a workflow's last remaining stage — delete the whole workflow with the delete_content_workflow tool instead of emptying it stage by stage. This cannot be undone. Like every workflow write, removing a stage recomputes the active stage (first stage with an open task), so an active stage that was removed is replaced automatically; `workflow.current_stage` never names the removed stage. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the ac…
-
add_content_workflow_taskAdd Content Workflow Task**Write operation.** Adds ONE new task to an existing stage of a workflow, leaving every other stage and task untouched — the safe alternative to the update_content_workflow tool when all you want is to insert a task. The task's title can be sent as `task` or `title` (see the schema). The target stage must already exist and must not already have a task with this title. A stage has at most 50 tasks — adding to a full one is rejected with a 400 on `task` ("Stage 'X' already has 50 tasks, the maximum."). Errors come back under the field they concern: `task` (title), `assignees`, `due_date`. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted. With a personal API key, a project whose setup is not finished can only be worked on by the user who c…
-
remove_content_workflow_taskRemove Content Workflow Task**Write operation.** Removes ONE task from a workflow stage, with its progress, assignments and comments. Every other stage and task is untouched. Refused on a stage's last remaining task — use the remove_content_workflow_stage tool to remove the whole stage instead of emptying it task by task. This cannot be undone. Does not consume credits. With a personal API key this endpoint needs the Content Suite full-access role (`suites.content_suite.role` is `full` in the get_account_me tool); any other role, for example Author, is refused with 403 (status_code 55024). Admins and the account-wide key are not restricted. With a personal API key, a project whose setup is not finished can only be worked on by the user who created it or by an admin; for anyone else the call answers 400 as for an unknown `project_id`. Removes the task and its progress/assignments/comments. Refused on a stage's last remaining task — use remove_content_workflow_stage instead. Confirm with the user first.
-
get_backlink_summaryGet Backlink SummaryReturns backlink totals, referring-domain counts and a spam score for the given target — a quick health check before a deeper backlink pull. Consumes Research Suite credits — 1 credit per call. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_backlinksGet BacklinksReturns currently live backlinks (source page, anchor text, dofollow/nofollow, rank) for the given target. Consumes Research Suite credits — 1 credit per call (up to 500 rows). A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Up to 500 rows cost 1 credit, but stick to the default 50 for an overview and raise it only when the user actually needs more rows.
-
audit_pageAudit PageCrawls one live page and returns an on-page optimization score plus title, description, word count, Core Web Vitals timing and SEO checks (missing tags, duplicate content, encoding issues, ...). Consumes Research Suite credits — 1 credit per call. Can take up to 60 seconds for a heavy page. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. One URL per call, can take up to 60 seconds for a heavy page — do not retry immediately on a timeout, ask the user first. This is the technical on-page check; for "how well can AI answers find and quote this page" (the AI Suite Page Audit, with prompt coverage) use create_ai_page_audit.
-
get_domain_overviewGet Domain OverviewReturns organic and paid keyword-position distributions (positions 1, 2-3, ..., 91-100), estimated traffic value and up/down/new/lost counts for a domain in a given market. Consumes Research Suite credits — 1 credit per call. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_demand_trendsGet Demand TrendsReturns a relative-popularity time series (0-100, normalized to the highest peak) for up to 5 keywords — useful for seasonality and demand-shift questions. The interval of the data points depends on `time_range` (weekly for long ranges). Consumes Research Suite credits — 1 credit per call. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
search_web_contentSearch Web ContentSearches a web-content index for pages mentioning a keyword (or an exact `"phrase"`), returning title, snippet, publish date and sentiment per page — useful for brand-mention and topic-coverage research beyond a single AI project. Consumes Research Suite credits — 1 credit per call. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
ask_llmAsk LlmSends a single prompt to an AI model (optionally with live web search) and returns its answer and any cited sources — for a one-off GEO check without setting up a tracked AI Suite project/prompt. Any available ChatGPT model can be chosen. Consumes Research Suite credits by the real price of the answer (model, answer length, web search), at least 1; `meta.credits_consumed` reports the charge. Before the call your balance must cover the upper estimate: 1 credit for small models (nano, mini, luna), 2 for all others, plus 1 with web search. Can take up to 60 seconds. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Sends the prompt to an external AI model, not Quaro's own data — use it only for an ad-hoc "how would an AI answer this" check, never as a substitute for Quaro's own tools. Can take up to 60 seconds.
-
search_llm_mentionsSearch Llm MentionsSearches across AI model answers for mentions of either a domain (your site or a competitor's) or a keyword/brand name, returning the matching question, the AI's answer and its cited sources — an ad-hoc GEO check without a tracked AI Suite project. Provide exactly one of `domain` or `keyword`. Consumes Research Suite credits by size - 2 credits up to 80 rows, 3 up to 170, 4 above; `meta.credits_consumed` reports the charge. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Provide exactly one of `domain` or `keyword`, never both and never neither.
-
get_backlink_anchorsGet Backlink AnchorsReturns the anchor texts used in backlinks to the target with backlink and referring-domain counts per anchor — shows how others describe the site. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_referring_domainsGet Referring DomainsReturns the referring domains of the target with backlink counts, rank and first/last seen dates. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_referring_networksGet Referring NetworksGroups the backlinks of the target by hosting network (IP address or subnet) — useful to spot link networks. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_backlink_competitorsGet Backlink CompetitorsReturns domains whose backlink profile overlaps with the target's (they are linked from the same referring domains) — link-building competitors. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
compare_backlink_profilesCompare Backlink ProfilesCompares the backlink profiles of 2-20 domains and returns the referring domains that link to all of them — link sources your competitors share. With exclude_targets (e.g. your own domain) referring domains that already link to those domains are dropped, which leaves the link gap. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Needs at least two different domains in `targets`; `exclude_targets` drops referring domains that also link to those domains (e.g. your own site).
-
compare_page_backlinksCompare Page BacklinksReturns the pages that link to all of the given 2-20 targets (pages or domains) — page-level link opportunities. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_domain_pagesGet Domain PagesReturns the individual pages of the target domain with their backlink and referring-domain counts. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_domain_pages_summaryGet Domain Pages SummaryReturns one summary row per page of the domain (backlinks, referring domains, rank, spam score) — find the strongest and the most linked pages. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_backlink_historyGet Backlink HistoryReturns backlink, referring-domain and rank totals of the target per day/week/month/year between two dates. Consumes Research Suite credits — 1 credit per call (2 for a very long daily series of more than about 1800 data points). `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_backlink_new_lost_historyGet Backlink New Lost HistoryReturns how many backlinks and referring domains were gained and lost per day/week/month/year between two dates. Consumes Research Suite credits — 1 credit per call (2 for a very long daily series of more than about 1800 data points). `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_backlink_countsGet Bulk Backlink CountsReturns the total number of backlinks for each of up to 1000 targets in one call — compare many sites cheaply. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_referring_domain_countsGet Bulk Referring Domain CountsReturns the number of referring domains for each of up to 1000 targets in one call. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_domain_ranksGet Bulk Domain RanksReturns the authority rank of each of up to 1000 targets in one call (scale 0-1000 by default). Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_spam_scoresGet Bulk Spam ScoresReturns the backlink spam score (0-100) of each of up to 1000 targets in one call — screen link sources or sites. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_pages_summaryGet Bulk Pages SummaryReturns a backlink summary (backlinks, referring domains, rank, spam score) per target for up to 1000 targets in one call. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_new_lost_backlinksGet Bulk New Lost BacklinksReturns the number of new and lost backlinks of each of up to 1000 targets since a date. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_bulk_new_lost_referring_domainsGet Bulk New Lost Referring DomainsReturns the number of new and lost referring domains of each of up to 1000 targets since a date. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
list_backlink_filtersList Backlink FiltersFree lookup: returns the field names (with their types) that can be used in `filters` and `order_by` of the backlink list endpoints. Without `endpoint` it returns the available list names; with `endpoint` the fields of that list. The list name is the endpoint path with underscores: anchors, referring_domains, referring_networks, competitors, domain_intersection, page_intersection, domain_pages, domain_pages_summary (e.g. referring_domains for the get_referring_domains tool). Does not consume credits. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Call this before using `filters` or `order_by` with a field you are not sure about; pass `endpoint` for the fields of one list.
-
get_domain_historyGet Domain HistoryReturns a monthly history of a domain's Google rankings: keyword counts per position range, estimated traffic and traffic value — shows whether a domain gains or loses visibility. Consumes Research Suite credits — 2 credits per call. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_historical_serpGet Historical SerpReturns Google results pages of a keyword collected over time: per snapshot the top 20 ranking results (organic, paid, featured snippet, local pack) with domain, title and URL; other result blocks are only counted in `other_item_types`. Shows how the top results for a keyword changed. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_domain_competitorsGet Domain CompetitorsReturns the domains that rank for the same keywords as the target in Google, with the number of shared keywords and visibility metrics. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
compare_domainsCompare DomainsReturns the keywords both domains rank for (or, with intersections=false, the keywords target1 ranks for and target2 does not) with the positions of each — the content gap between two sites. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
compare_pagesCompare PagesReturns the keywords that two or more page URLs rank for together (intersect) or any of them ranks for (union) with the positions per page. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_serp_competitorsGet Serp CompetitorsReturns the domains that appear most often in the Google results for the given keywords (up to 200) with visibility metrics — who you compete with on a topic. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_subdomainsGet SubdomainsReturns the subdomains of the target domain that rank in Google with keyword counts and estimated traffic per subdomain. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
estimate_domain_trafficEstimate Domain TrafficReturns the estimated monthly organic traffic and keyword counts of up to 1000 domains, pages or subdomains in one call. Consumes Research Suite credits by size: 1 credit up to 650 targets, 2 credits up to 1000 targets. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
get_domain_technologiesGet Domain TechnologiesReturns the technologies detected on a domain (CMS, shop system, analytics, ad and tag tools, frameworks, hosting ...) grouped by category. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024.
-
search_whoisSearch WhoisReturns registration data of domains: registrar, creation, change and expiry dates, registration status codes and TLD, plus organic/paid ranking metrics and backlink totals. `filters` is required and selects the domains you want, e.g. [["domain", "like", "%quaro%"]]. Consumes Research Suite credits by size: 2 credits up to 50 rows, 3 credits up to 125 rows, 4 credits up to 200 rows, 5 credits up to 275 rows, 6 credits up to 350 rows, 7 credits up to 425 rows, 8 credits up to 500 rows. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Select domains with `filters`, e.g. [["domain", "like", "%quaro%"]]; call list_whois_filters first for the available fields.
-
list_domain_filtersList Domain FiltersFree lookup: returns the field names (with their types) that can be used in `filters` and `order_by` of the domain analysis endpoints. Without `endpoint` it returns the available list names; with `endpoint` the fields of that list. List names: competitors_domain (the get_domain_competitors tool), domain_intersection (the compare_domains tool), page_intersection (the compare_pages tool), serp_competitors (the get_serp_competitors tool), subdomains (the get_subdomains tool). Does not consume credits. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Call this before using `filters` or `order_by` with a field you are not sure about; pass `endpoint` for the fields of one list.
-
list_technology_filtersList Technology FiltersFree lookup, for reference only: returns the field names (with their types) of the website-technology data. the get_domain_technologies tool itself takes no `filters`. Without `endpoint` it returns the available list names; with `endpoint` the fields of that list. Does not consume credits. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. For reference only - get_domain_technologies takes no filters. Pass `endpoint` for the fields of one list.
-
list_whois_filtersList Whois FiltersFree lookup: returns the field names (with their types) that can be used in `filters` and `order_by` of the search_whois tool. Without `endpoint` it returns the available list names; with `endpoint` the fields of that list. Does not consume credits. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Pass `endpoint` for the fields of one list.
-
extract_page_contentExtract Page ContentFetches a page and returns its content split into headings, paragraphs, lists, tables and images with the page's meta data — the text and structure a search engine reads. One URL per call, up to 60 seconds. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. One URL per call, can take up to 60 seconds - do not retry immediately on a timeout, ask the user first.
-
run_lighthouse_auditRun Lighthouse AuditRuns a Lighthouse audit (performance, accessibility, best practices, SEO) for one URL and returns the category scores, the core web vitals and the failed audits (the 25 with the lowest scores; `failed_audits_total` is the full count). Takes 30-90 seconds. Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Run it only when the user asks for Lighthouse, Core Web Vitals or performance scores; for a quick check use audit_page. Do not retry immediately on a timeout, ask the user first.
-
get_live_serpGet Live SerpFetches the current Google results page for a keyword in a location and language: organic results, featured snippets, People-also-ask and other result blocks with positions. Ad-hoc check — for keywords you track continuously, use a monitoring project (the get_project_rankings tool). Consumes Research Suite credits — normally 1 credit per call; an unusually large answer can cost more. `meta.credits_consumed` reports the credits charged. A query that matches nothing returns 200 with `data: null` and `meta.no_data: true` and is charged like a result. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Resolve the numeric `location_code` first (list_serp_locations, or list_google_searchvolume_locations). For tracked keywords use the monitoring tools instead - this is an ad-hoc check.
-
list_serp_locationsList Serp LocationsFree lookup: returns the locations (countries, states or regions, cities) available for live search-result checks in a country, with their numeric location_code. Narrow a long list with `query` (part of the name) or `location_type`. At most 500 locations are returned; `note` tells you when the list was cut. Does not consume credits. Personal keys need the Ideas right (right id 8); without it the call is refused with 403, status_code 55024. Use `query` (part of the name) or `location_type` to narrow long lists instead of reading all of them.
-
list_gsc_propertiesList Gsc PropertiesReturns the Search Console properties (domain properties like `sc-domain:example.com` and URL-prefix properties like `https://www.example.com/`) that an admin of the account has enabled for Quaro. Call this first and use a returned `site_url` for the other Search Console endpoints. An empty list means Search Console is connected but no property is enabled yet. This is your own Google data, not an estimate. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
get_gsc_performanceGet Gsc PerformanceReturns your own Google Search Console search-analytics data for one enabled property — clicks, impressions, CTR (in percent) and average position — broken down by up to three dimensions (query, page, country, device, date, searchAppearance), sorted by clicks. Use it for questions like "which queries bring traffic", "how did clicks develop", "which pages lost impressions"; compare two periods by calling it twice. Data covers at most the last 16 months and is complete up to about 3 days ago (the default end date). `rows` is capped at 1000 per call — page through with `start_row`. These are Google's real numbers for your site, not estimates. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
inspect_gsc_urlInspect Gsc UrlAsks Google Search Console how it sees one URL of an enabled property — indexed or not, why, last crawl, canonical chosen by Google, mobile usability and rich-result status. The URL must belong to the given property. Google limits this to about 2000 inspections per property and day, so use it for targeted checks, not for bulk crawling. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
list_gsc_sitemapsList Gsc SitemapsReturns the sitemaps submitted for an enabled Search Console property with their last submit/download time, pending state and the number of errors and warnings Google found. Use it to check whether a sitemap is registered and healthy. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
get_gsc_statusGet Gsc StatusReturns whether Google Search Console is connected for this account, whether the connection is still valid and which properties an admin enabled for Quaro (`allowed_properties`). `can_manage` is true when the current key may connect Google, enable properties or disconnect (a personal key of an account admin). Use it to check whether Search Console is set up before you call the other Search Console endpoints. The Google account's email address is never returned. Any key may read it. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
get_gsc_connect_linkGet Gsc Connect LinkAdmin only (the personal API key or MCP connection of an account admin; otherwise 403 / status_code 55024). Creates a one-time sign-in link that connects Google Search Console for the whole account. Open the returned `start_url` in a browser within 2 minutes (in an integration, hand it to the admin as a link), sign in with the Google account that has access to the Search Console properties and accept the read-only permission. The connection then belongs to the whole account (all users, Quinn (the AI chat in Quaro), the API and the MCP). Afterwards choose the properties Quaro may read (list them with the list_gsc_available_properties tool, then enable them with the set_gsc_enabled_properties tool). Also use it to reconnect when the connection expired. Quaro never needs the Google password, so never ask for it or send it. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
list_gsc_available_propertiesList Gsc Available PropertiesAdmin only (the personal API key or MCP connection of an account admin; otherwise 403 / status_code 55024). Lists ALL Search Console properties the connected Google account can see — not only the enabled ones — each with `allowed` (enabled for Quaro) and `suggested` (matches one of the `domains` you send; without `domains` nothing is marked). Use it to choose which properties Quaro may read, then enable them with the set_gsc_enabled_properties tool. Optional `domains`: your own domains, used only to mark `suggested`. To list only the enabled properties use the list_gsc_properties tool (works with any key). Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
set_gsc_enabled_propertiesSet Gsc Enabled PropertiesAdmin only (the personal API key or MCP connection of an account admin; otherwise 403 / status_code 55024). Sets which Search Console properties Quaro may read for the whole account (Quinn, API and MCP). The list REPLACES the current one: send every property that should stay enabled; an empty list enables none. Use the properties returned by the list_gsc_available_properties tool. Returns the new status. Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
-
disconnect_gscDisconnect GscAdmin only (the personal API key or MCP connection of an account admin; otherwise 403 / status_code 55024). Removes the Google Search Console connection of the whole account: all users, Quinn (the AI chat in Quaro), the API and the MCP lose access until an admin connects Google again (the get_gsc_connect_link tool) — the disconnect itself cannot be undone. Send `confirm: true` only after the account admin has given explicit confirmation (without `confirm: true` nothing happens). Does not consume credits. The rate limit is 30 requests per 60 seconds per API key, shared by all Search Console endpoints.
Connect your AI to Quaro
Add the official Quaro MCP server to Claude, ChatGPT or your editor.
Connect Quaro MCP to Claude
Claude.ai and Claude Desktop
Open Customize, then Connectors.
Click +, then Add custom connector.
Name it Quaro, paste the address below and click Add.
Connect Quaro MCP to ChatGPT
Open Settings, then Apps, then Advanced settings.
Turn on Developer mode, then click Create.
Name it Quaro and paste the address below.
ChatGPT usually only connects to servers with self-serve sign-in or none.
Connect Quaro MCP to Cursor
Opens Cursor if it's installed. If nothing happens, use the JSON below.
Connect Quaro MCP to VS Code
Opens VS Code if it's installed. Otherwise add the JSON settings from the Cursor tab to your .vscode/mcp.json.
Quaro MCP server questions
Does Quaro have an official MCP server?
Yes. Quaro publishes its MCP server itself. Its address is on Quaro's own domain (mcp.quaro.io). MartechDB only lists servers a vendor publishes, never community copies.
What is the Quaro MCP server URL?
The address is https://mcp.quaro.io/mcp.
How do I connect Quaro MCP to Claude, ChatGPT or Cursor?
In Claude, add https://mcp.quaro.io/mcp as a custom connector (Customize, Connectors, Add custom connector). In Cursor, use the Add to Cursor link on MartechDB's page, or paste the JSON settings into your mcp.json. In Claude Code, run: claude mcp add --transport http quaro https://mcp.quaro.io/mcp
What can the Quaro MCP server do?
It offers 206 tools, including delete_objects, get_account_credits, get_account_me, list_account_invoices and get_search_volume_status (as listed on 22 Nov 2026). 128 are marked read-only.
Is the Quaro MCP server in the official MCP Registry?
No. Quaro hasn't published it to the official MCP Registry.
Is the Quaro MCP server working?
Yes. It answered our most recent check. We last checked it on 22 Nov 2026.
What are the alternatives to Quaro's MCP server?
In SEO, the most agent-friendly alternatives with an official MCP server are Sitebulb (100/100), DataForSEO (100/100), Seolyze (97/100), Rebrandly (95/100) and SE Ranking (94/100). Quaro scores 55/100 on the same index.
Alternatives to Quaro's MCP server
Vendors in SEO with an official MCP server, the most agent-friendly first. Quaro scores 55/100.
Sitebulb
sitebulb.com
DataForSEO
dataforseo.com
Seolyze
seolyze.com
Rebrandly
rebrandly.com
SE Ranking
seranking.com
Change history
What we've seen change for Quaro, newest first.
-
Quaro: Quaro MCP server: 44 tools added (now 206)
-
Quaro: Quaro MCP server: 9 tools added (now 162)
-
Quaro: Quaro MCP server: 50 tools added (now 153)
Use with AI
Connect your AI to MartechDB
Then ask it about Quaro, or any of the 1,850 marketing tools with an official MCP server. Free, no sign-in.
Claude.ai and Claude Desktop
Open Customize, then Connectors.
Click +, then Add custom connector.
Name it MartechDB, paste the address below and click Add. There's no sign-in.
The free plan allows one custom connector. On Team and Enterprise plans, an owner adds the connector first.
Claude Code
Run this in your terminal:
ChatGPT
Open Settings, then Apps, then Advanced settings.
Turn on Developer mode, then click Create.
Name it MartechDB, paste the address below and choose No authentication.
Developer mode is on Plus and Pro plans, and on Business, Enterprise and Edu if an admin allows it.
API
The same data as JSON, with no sign-in. See the OpenAPI description for every endpoint and field.
CLI
Search the directory and print the command to connect a vendor's server, from your terminal. Needs Node 18 or later. View on npm
Show that your MCP server is verified
If you work at Quaro, add this badge to your MCP docs or README. It's free, it's the same for every vendor, and it updates from our daily checks: it shows uptime while the server responds, and says so if it stops.
Something wrong or out of date? Suggest a correction or email us, and we'll check it against the evidence. Every correction we make is listed on Corrections.