Account MCP for Search Console, audits, AI visibility and fixes. OAuth sign-in.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Without siteId: a read-only three-check probe of any public URL (favicon, canonical self-reference with hreflang x-default, BreadcrumbList). With siteId: run that site's full applicable rule set on this one page and diff every rule against the site's most recent completed audit run, reporting fixed / regressed / still failing — use this after you change a page. Read-only on both paths: it never writes an audit run or flips finding state, so a full audit is still what commits a resolution. Requires the audit:read scope.
Public http(s) URL to audit
Optional rule ids to check (default: every rule that applies to this page). Rule ids that do not apply are reported in notApplicable.
Optional. Scope the check to a site you can read: uses its rule set and compares against its latest completed run. The URL must belong to that site's domain.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http public https://visdenda.com/v1/mcp
Without siteId: a read-only three-check probe of any public URL (favicon, canonical self-reference with hreflang x-default, BreadcrumbList). With siteId: run that site's full applicable rule set on this one page and diff every rule against the site's most recent completed audit run, reporting fixed / regressed / still failing — use this after you change a page. Read-only on both paths: it never writes an audit run or flips finding state, so a full audit is still what commits a resolution. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| url* | string | Public http(s) URL to audit |
| rules | array | Optional rule ids to check (default: every rule that applies to this page). Rule ids that do not apply are reported in notApplicable. |
| siteId | string | Optional. Scope the check to a site you can read: uses its rule set and compares against its latest completed run. The URL must belong to that site's domain. |
{
"type": "object",
"required": [
"url"
],
"properties": {
"url": {
"type": "string",
"format": "uri",
"description": "Public http(s) URL to audit"
},
"rules": {
"type": "array",
"items": {
"type": "string"
},
"maxItems": 50,
"description": "Optional rule ids to check (default: every rule that applies to this page). Rule ids that do not apply are reported in notApplicable."
},
"siteId": {
"type": "string",
"description": "Optional. Scope the check to a site you can read: uses its rule set and compares against its latest completed run. The URL must belong to that site's domain."
}
},
"additionalProperties": false
}Check whether Google shows an AI Overview for up to 50 tracked terms and whether the site is referenced. Terms already sampled today for the same site are reused for free; missing terms trigger a live AI Overview query. Paid tool: dry-run without confirm:true; confirm:true needs an idempotencyKey. Requires the research:write scope.
| Argument | Type | Description |
|---|---|---|
| terms* | array | |
| locale | string | |
| siteId* | string | |
| confirm | boolean | |
| maxCostUsd | number | |
| maxCredits | integer | |
| idempotencyKey | string | |
| usePlatformCredits | boolean |
{
"type": "object",
"required": [
"siteId",
"terms"
],
"properties": {
"terms": {
"type": "array",
"items": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"maxItems": 50,
"minItems": 1
},
"locale": {
"type": "string"
},
"siteId": {
"type": "string"
},
"confirm": {
"type": "boolean",
"default": false
},
"maxCostUsd": {
"type": "number",
"default": 5,
"maximum": 50,
"exclusiveMinimum": 0
},
"maxCredits": {
"type": "integer",
"minimum": 1
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
},
"usePlatformCredits": {
"type": "boolean",
"default": false
}
},
"additionalProperties": false
}Run the site's applicable audit rules on 1–5 URLs that belong to that site's domain, and return per-rule verdicts (pass / warn / fail / na) with the checker detail for each URL. Narrow the set with rule ids when you only care about a few checks. Read-only: it never writes an audit run or flips finding state, and it does not diff against history — use audit_url for the before/after diff. Requires the audit:read scope and counts one network fetch per URL.
| Argument | Type | Description |
|---|---|---|
| urls* | array | 1–5 http(s) URLs on the site's domain. Duplicates are checked once. |
| rules | array | Optional rule ids to check (default: every rule that applies to the page). Ids that do not apply come back in notApplicable. |
| siteId* | string | Site under this API key; supplies the rule set and the allowed domain. |
{
"type": "object",
"required": [
"siteId",
"urls"
],
"properties": {
"urls": {
"type": "array",
"items": {
"type": "string",
"format": "uri"
},
"maxItems": 5,
"minItems": 1,
"description": "1–5 http(s) URLs on the site's domain. Duplicates are checked once."
},
"rules": {
"type": "array",
"items": {
"type": "string"
},
"maxItems": 50,
"description": "Optional rule ids to check (default: every rule that applies to the page). Ids that do not apply come back in notApplicable."
},
"siteId": {
"type": "string",
"description": "Site under this API key; supplies the rule set and the allowed domain."
}
},
"additionalProperties": false
}Use this after you deploy a fix: diff two committed runs of the same site — added/fixed/regressed, per-run pass/fail totals, truncated change list. Both runs must already exist; for a single page that has not been re-audited, use audit_url with siteId instead (read-only, no audit run consumed). Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string | Site id from list_sites |
| target* | string | Newer run id (target) |
| baseline* | string | Older run id (baseline) |
{
"type": "object",
"required": [
"siteId",
"baseline",
"target"
],
"properties": {
"limit": {
"type": "integer",
"default": 50,
"maximum": 200,
"minimum": 1
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
},
"target": {
"type": "string",
"description": "Newer run id (target)"
},
"baseline": {
"type": "string",
"description": "Older run id (baseline)"
}
},
"additionalProperties": false
}Rank what to fix from a site's latest run: failing rules grouped by page count, each linked to a P0/P1 radar update when the rule came from one. No writes, no network. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 10,
"maximum": 20,
"minimum": 1
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Return cache hits, billable items, estimated provider cost and whether the call fits maxCostUsd for one of the four paid research tools, using the same parameters you would pass to it. Zero vendor calls and zero writes. Requires the research:read scope.
| Argument | Type | Description |
|---|---|---|
| tool* | string | |
| depth | integer | |
| limit | integer | |
| seeds | array | |
| terms | array | |
| device | string | |
| locale | string | |
| siteId | string | |
| queries | array | |
| maxCostUsd | number | |
| maxCredits | integer | |
| usePlatformCredits | boolean |
{
"type": "object",
"required": [
"tool"
],
"properties": {
"tool": {
"enum": [
"get_serp_results",
"get_keyword_metrics",
"research_keywords",
"check_ai_visibility"
],
"type": "string"
},
"depth": {
"type": "integer"
},
"limit": {
"type": "integer"
},
"seeds": {
"type": "array",
"items": {
"type": "string"
}
},
"terms": {
"type": "array",
"items": {
"type": "string"
}
},
"device": {
"enum": [
"desktop",
"mobile"
],
"type": "string"
},
"locale": {
"type": "string"
},
"siteId": {
"type": "string"
},
"queries": {
"type": "array",
"items": {
"type": "string"
}
},
"maxCostUsd": {
"type": "number",
"maximum": 50,
"exclusiveMinimum": 0
},
"maxCredits": {
"type": "integer",
"minimum": 1
},
"usePlatformCredits": {
"type": "boolean"
}
},
"additionalProperties": false
}Explain a single finding: the page, status and detail, plus the rule's title, category, documentation link and verification guidance as stored. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| runId* | string | Audit run id |
| findingId* | string | Finding id within that run |
{
"type": "object",
"required": [
"runId",
"findingId"
],
"properties": {
"runId": {
"type": "string",
"description": "Audit run id"
},
"findingId": {
"type": "string",
"description": "Finding id within that run"
}
},
"additionalProperties": false
}Fetch one public HTML page and return what is actually on it: final URL, status code, redirect count, title, meta description, canonical, meta robots, html lang, H1 texts, heading counts, JSON-LD @types, word count and HTML size. No rule verdicts — call check_url for those, or audit_url to diff against the site's last completed run. Use it to look before you decide, or to see what a change actually shipped. Requires the audit:read scope and counts one network fetch.
| Argument | Type | Description |
|---|---|---|
| url* | string | Public http(s) URL to fetch. Private hosts and non-HTML responses are rejected. |
{
"type": "object",
"required": [
"url"
],
"properties": {
"url": {
"type": "string",
"format": "uri",
"description": "Public http(s) URL to fetch. Private hosts and non-HTML responses are rejected."
}
},
"additionalProperties": false
}Search internal anchor texts in the latest audit crawl snapshot: returns matching anchor texts, how many times each occurs, which pages use it (up to 20) and which pages it points to (up to 20). Use it to audit generic or over-optimised anchor text, or to find pages that already link to a target. An empty anchor argument returns the most used anchor texts. Exact matches rank before substring matches. The result carries the snapshot coverage and whether the anchor index was truncated, so a partial crawl or capped index is never read as the whole site. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | Maximum anchor texts to return (default 20). |
| anchor | string | Anchor text to search for (case-insensitive substring; exact matches rank first). Omit to list the most used anchors. |
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"maximum": 100,
"minimum": 1,
"description": "Maximum anchor texts to return (default 20)."
},
"anchor": {
"type": "string",
"maxLength": 200,
"description": "Anchor text to search for (case-insensitive substring; exact matches rank first). Omit to list the most used anchors."
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Read what AI crawlers and AI referrals actually did on a site in the last 7, 28 or 30 days: hits by agent, previous-window delta, markdown negotiation, machine-file reads, AI referral hosts, and robots violations. Use this when the user asks whether AI bots visit, whether they are blocked, or whether AI sends traffic. Returns na with a reason when the site has no probe data; missing data is never treated as zero. Read-only. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
30
],
"type": "integer",
"default": 28
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read the latest AI Overview check for the site's tracked queries: whether an AI Overview appeared and whether it cited the user's domain. Paid plans also get up to 90 days of sampled presence for up to 10 queries. Use this when the user asks whether they appear in Google AI Overviews or whether AI answers cite them. An empty trend means not sampled, not absent. Read-only. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| query | string | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
90
],
"type": "integer",
"default": 28
},
"query": {
"type": "string"
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read findings for a site (latest run by default, or one run by id) with server-side filters: status, severity, rule and page substring. Returns rule, page, status, detail, evidence key plus facet counts so the caller can see what else is in the run. Rule text is returned as authored. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| page | string | Only return findings whose page URL contains this substring |
| limit | integer | |
| runId | string | Audit run id; defaults to the latest run |
| offset | integer | Skip this many findings (pagination) |
| ruleId | string | Only return findings for this rule |
| siteId* | string | Site id from list_sites |
| status | string | Only return findings with this status |
| severity | string | Only return findings frozen at this severity |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"page": {
"type": "string",
"description": "Only return findings whose page URL contains this substring"
},
"limit": {
"type": "integer",
"default": 20,
"maximum": 100,
"minimum": 1
},
"runId": {
"type": "string",
"description": "Audit run id; defaults to the latest run"
},
"offset": {
"type": "integer",
"default": 0,
"maximum": 100000,
"minimum": 0,
"description": "Skip this many findings (pagination)"
},
"ruleId": {
"type": "string",
"description": "Only return findings for this rule"
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
},
"status": {
"enum": [
"fail",
"warn",
"pass",
"na"
],
"type": "string",
"description": "Only return findings with this status"
},
"severity": {
"enum": [
"blocker",
"warning",
"info"
],
"type": "string",
"description": "Only return findings frozen at this severity"
}
},
"additionalProperties": false
}A site's recent audit runs, newest first, with status, summary, coverage quality and trigger. Use it to pick a run, or to judge improvement across runs. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 5,
"maximum": 20,
"minimum": 1
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Pass rate and added/fixed/regressed across a site's last N runs — the 'is it getting better' view rather than just the latest run. Needs at least two completed runs. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 10,
"maximum": 20,
"minimum": 1
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Read the deterministic 30-day brief for every competitor under one own site: new page counts, coverage against your pages, recurring title themes, top gaps and publishing pace (last 7 days vs the 7 before). Includes the cached AI explanation when one was generated. Requires the sites:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read recent sitemap/content changes for one competitor target; use get_competitor_feed for page text and coverage. New, removed, lastmod-updated and content-changed pages in the last N days (default 7). Requires the sites:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| type | string | |
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"type": "integer",
"default": 7,
"maximum": 30,
"minimum": 1
},
"type": {
"enum": [
"added",
"removed",
"lastmod_changed",
"content_changed",
"fetch_failed"
],
"type": "string"
},
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}List candidate competitors that appear in the top 10 of your tracked keywords (from the latest keyword-compare evidence): appearances, how many keywords you both rank for, average position and sample queries. Dismissed domains are excluded and returned separately. Requires the sites:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read one competitor's detected page feed with covered/partial/gap matching and excerpts; filter by match or triage. Pages whose content has not been fetched yet appear with matchReason=not-enriched and are counted in counts.pending. Matching compares page content signatures against the parent site's latest completed audit pages, never against Google or traffic data. Requires the sites:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| match | string | |
| siteId* | string | |
| triage | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"match": {
"enum": [
"covered",
"partial",
"gap",
"na"
],
"type": "string"
},
"siteId": {
"type": "string"
},
"triage": {
"enum": [
"new",
"saved",
"dismissed"
],
"type": "string"
}
},
"additionalProperties": false
}List one competitor's pages with no equivalent own-site coverage (match=gap), lowest similarity first. Each row carries the closest own pages so you can decide whether to write, merge or ignore. Requires the sites:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read the latest keyword-compare run for one own site: top-10 positions per keyword for your site and each observed competitor, capture status and classification. SERP evidence only; no traffic estimates. Requires the sites:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Paths AI and search crawlers requested that returned 404/410, paths blocked by robots or 403, and whether the machine-facing files (robots.txt, llms.txt, sitemap.xml, markdown) were fetched at all. This is the only signal that turns 'what machines wanted' into a task: create, redirect, unblock or publish llms.txt. Read-only. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
30
],
"type": "integer",
"default": 28
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}The open finding backlog with age buckets, resolutions in the last 30 days and the rules with the most open items — use it to judge whether the backlog is ageing or being cleared. Needs more than one completed run to mean anything. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}One call over the live GSC query×page matrix: queries where two or more qualifying pages (at least 2 impressions, at least 10% of the query's impressions, position 30 or better) compete, ordered by clicks landing on non-leader pages. Each group carries the leader's impression share and each page's position and clicks, winner first — evidence for fold/canonical decisions, not a deletion verdict. Live GSC read on paid plans, daily-capped. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 10,
"maximum": 20,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Pages losing clicks between two equal windows (28 or 90 days, versus the previous window or the same window last year), each labeled by cause: decayed-content means position held and the content aged (refresh it), ranking-drop means position fell and a refresh will not fix it. Rows are ordered by clicks lost and capped; counts always cover the full evaluated set, steady rows are excluded from the list. Live GSC read on paid plans, daily-capped. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string | |
| compare | string | |
| windowDays | integer |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 15,
"maximum": 25,
"minimum": 1
},
"siteId": {
"type": "string"
},
"compare": {
"enum": [
"prev",
"yoy"
],
"type": "string",
"default": "prev"
},
"windowDays": {
"enum": [
28,
90
],
"type": "integer",
"default": 28
}
},
"additionalProperties": false
}One month of Google Search Console data (clicks, impressions, CTR, position) from the stored archive; falls back to a live read with the key owner's credentials (paid plans only, daily-capped). Search performance exists only in GSC — it cannot be derived from the site or its code. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| month | string | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"month": {
"type": "string"
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Clicks, impressions, CTR and average position for the last 7, 28 or 90 days with previous-period deltas. Live GSC data (final data lags ~3 days); paid plans, counts against the daily live-read cap. This is the only source for search performance — nothing in the site's code or HTML reveals it. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | Window length ending ~3 days ago |
| siteId* | string | |
| includeDaily | boolean | Include the day-by-day series |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
90
],
"type": "integer",
"default": 28,
"description": "Window length ending ~3 days ago"
},
"siteId": {
"type": "string"
},
"includeDaily": {
"type": "boolean",
"default": false,
"description": "Include the day-by-day series"
}
},
"additionalProperties": false
}Top landing pages by clicks for the last 7, 28 or 90 days with impressions, CTR, position and previous-period deltas. Live GSC read on paid plans, daily-capped; not derivable from the site itself. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
90
],
"type": "integer",
"default": 28
},
"limit": {
"type": "integer",
"default": 10,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Top search queries by clicks for the last 7, 28 or 90 days with impressions, CTR, position and previous-period deltas. Live GSC read on paid plans, daily-capped; not derivable from the site itself. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
90
],
"type": "integer",
"default": 28
},
"limit": {
"type": "integer",
"default": 10,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Queries where the site ranks 4-20 with above-median impressions, each row carrying the page, position, impressions, CTR and potentialClicks — the click headroom if the page reached the site's own top-position CTR (a site baseline, not an industry curve), best opportunity first. The server does the ranking; do not re-rank by impressions alone. Live GSC read on paid plans, daily-capped; not derivable from the site itself. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"limit": {
"type": "integer",
"default": 10,
"maximum": 25,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}The last 7 days versus the previous 7 days in a single read: totals (clicks, impressions, CTR, position) with deltas, top-5 winning and losing pages, top-5 new and lost queries, top-3 quick wins and the top-3 pages with the largest position swing. Every list is bounded and server-ranked; numbers are co-occurrence observations and never name a cause — pair a drop with get_radar_updates before explaining it. Live GSC read on paid plans, daily-capped. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read volume, difficulty, CPC and competition for up to 1,000 terms that are already tracked for one site. Fresh rows reuse the 30-day metrics cache; only stale or missing terms hit DataForSEO. Never creates a report (reportId is always null). Paid tool: confirm:true plus idempotencyKey is required to execute. Requires the research:write scope.
| Argument | Type | Description |
|---|---|---|
| terms* | array | |
| locale | string | |
| siteId* | string | |
| confirm | boolean | |
| maxCostUsd | number | |
| maxCredits | integer | |
| idempotencyKey | string | |
| usePlatformCredits | boolean |
{
"type": "object",
"required": [
"siteId",
"terms"
],
"properties": {
"terms": {
"type": "array",
"items": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"maxItems": 1000,
"minItems": 1
},
"locale": {
"type": "string"
},
"siteId": {
"type": "string"
},
"confirm": {
"type": "boolean",
"default": false
},
"maxCostUsd": {
"type": "number",
"default": 5,
"maximum": 50,
"exclusiveMinimum": 0
},
"maxCredits": {
"type": "integer",
"minimum": 1
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
},
"usePlatformCredits": {
"type": "boolean",
"default": false
}
},
"additionalProperties": false
}Rank pages worth fixing: latest-run fails weighted by archived-month clicks. Empty means no evidence, not no problems. The weighting needs GSC clicks, so this ranking cannot be reproduced from the repo. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| month | string | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"month": {
"type": "string"
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Read the internal link neighbourhood of one page from the latest audit crawl snapshot: inlink and outlink counts with up to 20 example URLs each, crawl depth, URL template cluster, and whether the page is an orphan. Use it before recommending internal links or judging whether a page is reachable; a page that is not in the snapshot is reported with inSnapshot=false rather than an error. The result carries the snapshot coverage (crawled pages, sitemap URLs, truncated) so a partial crawl is never read as the whole site. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| url* | string | Page URL or path on the site's domain |
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId",
"url"
],
"properties": {
"url": {
"type": "string",
"description": "Page URL or path on the site's domain"
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Everything known about a single URL in one read: GSC queries/clicks for that page, sessions and engagement, AI crawler fetches, AI referral clicks, the latest index verdict, open audit findings and real-user CWV. Paid plans read GSC live for that page (queries and average position); otherwise the archived page snapshot is returned with search.source=archive and empty queries. The join is per page, so numbers are not duplicated across queries. Each block carries its own na/reason — a missing source is not a zero. Read-only. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| url* | string | |
| days | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId",
"url"
],
"properties": {
"url": {
"type": "string",
"maxLength": 2048
},
"days": {
"enum": [
7,
28,
30
],
"type": "integer",
"default": 28
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Every site this key can read with its latest run outcome (clear, attention, interrupted, etc.), open finding count and totals, attention-first. Use it to pick which site to inspect next. Prefer `outcome` over the coarse legacy `runStatus`. Requires the audit:read scope.
No arguments.
{
"type": "object",
"properties": {},
"additionalProperties": false
}Start here for one site's state: latest and previous audit runs with coverage, the delta since the previous run (new failures, resolved, pages added/removed), highest-priority fixes with severity and sample URLs, recent P0/P1 radar updates that hit its failing rules, uptime/CWV/watchdog signals, and the archived GSC month. The delta comes from stored audit history, which no source repo can produce. Read-only, no live GSC or page fetch. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | Site id from list_sites |
| radarDays | integer | Only radar updates published within this many days |
| priorityLimit | integer |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string",
"description": "Site id from list_sites"
},
"radarDays": {
"type": "integer",
"default": 14,
"maximum": 90,
"minimum": 1,
"description": "Only radar updates published within this many days"
},
"priorityLimit": {
"type": "integer",
"default": 5,
"maximum": 20,
"minimum": 1
}
},
"additionalProperties": false
}Recent P0/P1 Google Search updates and how many active audit rules each one links to. This is change outside the site and outside its source code — use it to explain why a rule matters now. Requires the radar:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | Max items to return (default 5, max 10) |
| since | string | Only updates published on or after this date (YYYY-MM-DD) |
{
"type": "object",
"properties": {
"limit": {
"type": "integer",
"default": 5,
"maximum": 10,
"minimum": 1,
"description": "Max items to return (default 5, max 10)"
},
"since": {
"type": "string",
"format": "date",
"description": "Only updates published on or after this date (YYYY-MM-DD)"
}
},
"additionalProperties": false
}Read one saved research report: SERP results, keyword ideas or AI Overview checks plus the current GSC query rows, AI Overview samples and tracked keywords for the same terms. Missing context is returned as status na with a reason. Reads only; never calls DataForSEO and never creates a new run. Requires the research:read scope.
| Argument | Type | Description |
|---|---|---|
| reportId* | string |
{
"type": "object",
"required": [
"reportId"
],
"properties": {
"reportId": {
"type": "string"
}
},
"additionalProperties": false
}Return one rule's definition: what it checks, severity, requirement, scope, which pages and site types it applies to, the official source link and version, and whether the rule is active, draft or retired (with the retirement reason). Use it to explain a finding, to cite the rule's source, or to decide which rule to check next. Requires the rules:read scope; reads only the database.
| Argument | Type | Description |
|---|---|---|
| ruleId* | string | Rule id as returned by get_audit_findings, check_url or audit_url. |
{
"type": "object",
"required": [
"ruleId"
],
"properties": {
"ruleId": {
"type": "string",
"description": "Rule id as returned by get_audit_findings, check_url or audit_url."
}
},
"additionalProperties": false
}Run one live Google SERP read per query (no 3-read merge, no averaging) and return ranked organic results plus SERP extras from the same response: feature types present, People Also Ask questions, related searches and the featured snippet. Saved as a research report. Paid tool: without confirm:true it only returns a dry-run estimate with cache hits; confirm:true requires an idempotencyKey and respects maxCostUsd. Cached for 12 hours per account. Requires the research:write scope.
| Argument | Type | Description |
|---|---|---|
| depth | integer | |
| device | string | |
| locale | string | |
| confirm | boolean | |
| queries* | array | |
| maxCostUsd | number | |
| maxCredits | integer | |
| idempotencyKey | string | |
| usePlatformCredits | boolean |
{
"type": "object",
"required": [
"queries"
],
"properties": {
"depth": {
"type": "integer",
"default": 10,
"maximum": 100,
"minimum": 10,
"multipleOf": 10
},
"device": {
"enum": [
"desktop",
"mobile"
],
"type": "string",
"default": "desktop"
},
"locale": {
"enum": [
"us",
"all"
],
"type": "string",
"default": "us"
},
"confirm": {
"type": "boolean",
"default": false
},
"queries": {
"type": "array",
"items": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"maxItems": 10,
"minItems": 1
},
"maxCostUsd": {
"type": "number",
"default": 5,
"maximum": 50,
"exclusiveMinimum": 0
},
"maxCredits": {
"type": "integer",
"minimum": 1
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
},
"usePlatformCredits": {
"type": "boolean",
"default": false
}
},
"additionalProperties": false
}Technology fingerprint (framework, router, rendering, i18n lib, hosting signals), from the 24h cache when fresh. If you can read the site's source repository, prefer the repo — it is authoritative and richer; use this for a site whose code you do not have. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}The site's 0–100 health score with four explainable penalties (open findings, backlog age, audit freshness, coverage) and the raw inputs behind them. Computed from audit history, so a source repo cannot produce it; null means no completed audit yet. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Internationalization maturity level and gaps, derived from the tech fingerprint. If you can read the source repository, prefer the repo; use this for a site whose code you do not have. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Start here before raw reads: deterministic insights across GSC, audit, indexing and AI-crawler data, each with typed evidence, severity, impact and a suggested next action. The list is ranked and says what could not be checked and why; an empty list means no rule fired, not that the site is healthy. Live GSC is used only when the plan and daily budget allow; otherwise the GSC types are reported as not evaluated and the other sources still return. Read-only. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | Comparison window; GSC final data ends ~3 days ago |
| limit | integer | |
| types | array | Optional subset of insight types; omit for all |
| siteId* | string | Site id from list_sites |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
90
],
"type": "integer",
"default": 28,
"description": "Comparison window; GSC final data ends ~3 days ago"
},
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"types": {
"type": "array",
"items": {
"enum": [
"declining_page",
"high_impressions_low_ctr",
"cannibalization",
"audit_fail_on_traffic_page",
"index_coverage_gap",
"crawler_gap",
"crawled_not_cited",
"high_exit_low_read",
"page_performance_regression"
],
"type": "string"
},
"maxItems": 9,
"description": "Optional subset of insight types; omit for all"
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
}
},
"additionalProperties": false
}Per-page join of GSC clicks and average position (aggregated by page, never per query), sessions and engaged rate, AI crawler fetches, AI referral sessions, index verdict, open findings and LCP. Search numbers are read live per page when the plan and daily budget allow; otherwise they fall back to the archived snapshot and position stays empty. Each row and the response say which source was used. Use it to find pages with traffic but no engagement, or pages AI reads that no human visits. Empty means no evidence, not no problems. Read-only. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| days | integer | |
| limit | integer | |
| siteId* | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"days": {
"enum": [
7,
28,
30
],
"type": "integer",
"default": 28
},
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Given a radar update and a site, return the active rules linked to that update plus that site's latest-run findings for those rules. It joins an external Google change to your own failures — neither side alone tells you whether the update hits you. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | Site id from list_sites |
| updateId* | string | Radar update id from get_radar_updates |
{
"type": "object",
"required": [
"updateId",
"siteId"
],
"properties": {
"siteId": {
"type": "string",
"description": "Site id from list_sites"
},
"updateId": {
"type": "string",
"description": "Radar update id from get_radar_updates"
}
},
"additionalProperties": false
}Fetch the site's sitemap and add new pages to its audit list, truncated at the plan's auditUrls limit. Requires the audits:write scope. Writes need confirm:true plus a unique idempotencyKey; without confirm:true the call returns a dry-run preview and changes nothing.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | |
| confirm | boolean | |
| idempotencyKey | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
},
"confirm": {
"type": "boolean",
"default": false
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
}
},
"additionalProperties": false
}List proposals owned by this key's account, newest first, with status counts. Defaults to pending; expired proposals are reported as expired after the server sweeps stale pending rows. Approved means the owner authorized the change (not that it was executed); applied means the owner marked the authorized change as executed; rejected proposals carry the owner's decision note. Requires the audit:read scope and a key owned by an account.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| status | string |
{
"type": "object",
"required": [],
"properties": {
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"status": {
"enum": [
"pending",
"approved",
"applied",
"rejected",
"withdrawn",
"expired"
],
"type": "string",
"default": "pending"
}
},
"additionalProperties": false
}List research reports for this account with kind, site, saved status, cost and reuse window. Reads only; never calls DataForSEO. Requires the research:read scope.
| Argument | Type | Description |
|---|---|---|
| kind | string | |
| limit | integer | |
| siteId | string |
{
"type": "object",
"required": [],
"properties": {
"kind": {
"enum": [
"serp",
"keywords",
"aio",
"mixed"
],
"type": "string"
},
"limit": {
"type": "integer",
"default": 20,
"maximum": 50,
"minimum": 1
},
"siteId": {
"type": "string"
}
},
"additionalProperties": false
}Start here: the sites this key can read, with verification state and the latest audit run. Requires the audit:read scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | Max sites to return (default 50, max 200) |
{
"type": "object",
"properties": {
"limit": {
"type": "integer",
"default": 50,
"maximum": 200,
"minimum": 1,
"description": "Max sites to return (default 50, max 200)"
}
},
"additionalProperties": false
}Write a pending change request instead of changing the site: title, summary, rationale, 1–20 concrete changes and optional rule ids. The site owner approves or rejects it in the console; rejection notes come back through list_proposals. Use this when the fix is a judgment call or touches the user's pages — it does not modify the site, write an audit run or consume an audit run. Requires the audits:write scope and a key owned by an account. Repeating an identical proposal while one is pending returns the existing row with deduplicated=true.
| Argument | Type | Description |
|---|---|---|
| title* | string | |
| siteId* | string | Site id from list_sites |
| changes* | array | |
| ruleIds | array | Optional rule ids this proposal addresses |
| summary* | string | One-paragraph statement of what should change and why it matters |
| rationale* | string | The evidence chain: findings, rules, pages and expected effect |
| targetUrl | string | Optional page URL; must be http(s) and on the site's domain |
{
"type": "object",
"required": [
"siteId",
"title",
"summary",
"rationale",
"changes"
],
"properties": {
"title": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"siteId": {
"type": "string",
"description": "Site id from list_sites"
},
"changes": {
"type": "array",
"items": {
"type": "object",
"required": [
"target",
"change"
],
"properties": {
"change": {
"type": "string",
"maxLength": 500,
"minLength": 1
},
"target": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"current": {
"type": [
"string",
"null"
],
"maxLength": 2000
},
"proposed": {
"type": [
"string",
"null"
],
"maxLength": 2000
}
},
"additionalProperties": false
},
"maxItems": 20,
"minItems": 1
},
"ruleIds": {
"type": "array",
"items": {
"type": "string"
},
"maxItems": 20,
"description": "Optional rule ids this proposal addresses"
},
"summary": {
"type": "string",
"maxLength": 2000,
"minLength": 1,
"description": "One-paragraph statement of what should change and why it matters"
},
"rationale": {
"type": "string",
"maxLength": 4000,
"minLength": 1,
"description": "The evidence chain: findings, rules, pages and expected effect"
},
"targetUrl": {
"type": "string",
"format": "uri",
"description": "Optional page URL; must be http(s) and on the site's domain"
}
},
"additionalProperties": false
}Ask DataForSEO Labs for keyword ideas from up to 5 seeds with limit 50/150/200 per seed, returning term, volume, difficulty and intent. Results are cached for 24 hours per account and saved as a research report. Paid tool: dry-run without confirm:true; confirm:true needs an idempotencyKey. Requires the research:write scope.
| Argument | Type | Description |
|---|---|---|
| limit | integer | |
| seeds* | array | |
| locale | string | |
| confirm | boolean | |
| maxCostUsd | number | |
| maxCredits | integer | |
| idempotencyKey | string | |
| usePlatformCredits | boolean |
{
"type": "object",
"required": [
"seeds"
],
"properties": {
"limit": {
"enum": [
50,
150,
200
],
"type": "integer",
"default": 150
},
"seeds": {
"type": "array",
"items": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"maxItems": 5,
"minItems": 1
},
"locale": {
"enum": [
"us",
"all"
],
"type": "string",
"default": "us"
},
"confirm": {
"type": "boolean",
"default": false
},
"maxCostUsd": {
"type": "number",
"default": 5,
"maximum": 50,
"exclusiveMinimum": 0
},
"maxCredits": {
"type": "integer",
"minimum": 1
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
},
"usePlatformCredits": {
"type": "boolean",
"default": false
}
},
"additionalProperties": false
}Start a full audit for one verified site owned by this API key. This is the call that commits finding state (resolved / regressed) and feeds trends and health. It reserves audit page checks from the account's monthly pool (settled by successfully fetched pages) and queues the run when the account or global audit capacity is busy; a queued result includes queuePosition and queueEtaSeconds. Requires the audits:write scope. Writes need confirm:true plus a unique idempotencyKey; without confirm:true it returns a dry-run preview and changes nothing. Reusing the same idempotencyKey within 24h returns the first result.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | |
| confirm | boolean | |
| idempotencyKey | string |
{
"type": "object",
"required": [
"siteId"
],
"properties": {
"siteId": {
"type": "string"
},
"confirm": {
"type": "boolean",
"default": false
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
}
},
"additionalProperties": false
}Run the console's single-page fix re-check and commit the result: a pass resolves the finding as fix-applied; anything else only records the verdict and leaves it open. Use this after changing a page, once audit_url shows the rule now passes. The result carries a research block with the latest SERP snapshot rank and AI Overview sample per query (up to 3; omitted queries default to the page's top Search Console queries — the block reads already-collected research data and never calls a vendor). It does not write an audit run or consume an audit run. Requires the audits:write scope. Writes need confirm:true plus a unique idempotencyKey; without confirm:true it returns a dry-run preview and changes nothing. Reusing the same idempotencyKey within 24h returns the first result.
| Argument | Type | Description |
|---|---|---|
| siteId* | string | |
| confirm | boolean | |
| queries | array | |
| findingStateId* | string | |
| idempotencyKey | string |
{
"type": "object",
"required": [
"siteId",
"findingStateId"
],
"properties": {
"siteId": {
"type": "string"
},
"confirm": {
"type": "boolean",
"default": false
},
"queries": {
"type": "array",
"items": {
"type": "string",
"maxLength": 200,
"minLength": 1
},
"maxItems": 3
},
"findingStateId": {
"type": "string"
},
"idempotencyKey": {
"type": "string",
"maxLength": 128,
"minLength": 1
}
},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 22, 2026.