Semrush MCP

Semrush is an SEO and competitive research platform, and its MCP server is a hosted remote endpoint that lets an AI client pull the same reports the web app and the public API serve. Instead of exporting spreadsheets and pasting them into a prompt, an assistant queries the data mid conversation and reasons over the rows it gets back. The server is organized around discovery tools for each report family, covering domain overview, organic research, keyword research, competitors research, backlinks research, paid search research, audience and traffic research, shopping research, position tracking, and site audit. The flow is deliberate: an agent picks a report, calls a schema tool to learn which parameters that report accepts, then executes it and receives structured results. Project data such as a site audit or a rank tracking campaign is available read only, so an assistant can analyze a project but cannot create or modify one. Access runs over OAuth against a Semrush account, with an API key header as a fallback, and the endpoint speaks streamable HTTP rather than a local stdio process. Requests draw on the account's API unit balance exactly as direct API calls do, and report availability follows the subscription. It fits SEO teams and content strategists who want keyword research, competitor gap analysis, and backlink checks done inside an assistant, and agencies wiring Semrush data into reporting agents.