Real search volume and CPC, not estimates
research_keywords
SEO MCP
Give your agent keyword, SERP, competitor, and site health data through one OpenRush connection.
Your agent sizes demand and maps the competition without you opening a dashboard.
research_keywords
discover_competitors
inspect_serp
Audit, track, and drill into the pages and keywords that matter.
audit_site
inspect_search_visibility
compare_keyword_coverage
Every tool the OpenRush SEO MCP exposes. Each one answers a question rather than returning a table for the agent to parse.
inspect_domainTop keywords, pages, traffic, and likely competitors for a domaindiscover_competitorsOrganic competitors by keyword overlapresearch_keywordsExpand a seed into ranked demand with volume, CPC, and intentcompare_keyword_coverageKeyword gaps against up to ten competitorsinspect_serpLive result page with AI overview, snippets, and paid resultsinspect_search_visibilityRanking positions across a keyword setinspect_keywordOne keyword in full: trend, SERP, and features it triggersinspect_pageKeywords and traffic for a single URLaudit_siteTechnical and content health over sampled pagesexport_datasetFull rows for any result too large to inlinedescribe_capabilitiesWhat OpenRush exposes, so the agent never guessesReturns the full upstream payload. One SERP call can run to tens of thousands of tokens of nested JSON, which is why some providers tell you to write it to a file before your agent reads it. Expose a hundred and sixty tools on top of that and the model spends its budget choosing between them.
A compact envelope holding the answer, typed facts each carrying their own source and confidence, the entities involved, what the result does not cover, and the next calls worth making. The agent gets something it can act on and cite, not something it has to parse.
Free SEO resources
Keywords
Ranked by average monthly global search volume. Snapshot taken June 2026.
Browse dataset
Websites
Ranked by how many Google keywords each domain ranks for, with estimated monthly traffic value alongside. Snapshot taken June 2026.
Browse dataset
Backlinks
Ranked by backlink rank, a 0 to 1000 score of how much of the web points at each domain. Snapshot taken June 2026.
Browse dataset
An SEO MCP lets your agent pull keyword, SERP, competitor, and site health data directly in conversation.
Any MCP-compatible client. Claude, ChatGPT, Claude Code, Codex, Cursor, and Perplexity all have a documented path on the setup page. Claude users can add it in one click from the connector directory.
No. Connecting OpenRush is a settings change in your agent, not an integration project: you paste one URL, or add it to Claude in a single click. The REST API exists for product integrations, but you do not need it to use OpenRush from a chat client.
A raw API wrapper returns whatever the API returns, which for a single SERP can be tens of thousands of tokens of nested JSON. OpenRush returns a compact envelope: the answer, the facts behind it with per-fact provenance and confidence, what the result does not cover, and suggested next calls. Eleven tools rather than a hundred also means the agent picks the right one.
Live Google result pages, a search index covering 8 billion keywords and 1.3 billion websites refreshed daily, a backlink index, and direct crawls for site audits. Every fact carries its own source class and observation timestamp, so the agent can tell a live SERP reading apart from an indexed estimate.
One-click setup with Claude or any MCP-compatible agent.