Semrush MCP Server

Semrush

Semrush is an SEO and paid-search research platform. Marketers use it for domain visibility, keyword discovery, competitive gaps, backlink profiles, and technical site audits on tracked projects.

First moves

1. API units balance — check remaining credits before a multi-report session.
2. Domain overview — authority, traffic, and keyword counts for one market or worldwide.
3. Keyword research — phrase overview, related terms, question-style keywords, or difficulty for a target phrase.
4. Rank history — visibility trend for a domain in a chosen regional database.
5. Content gaps — organic or paid keyword lists and competitor overlap for a domain.
6. URL-level rankings — organic keywords ranking for a specific page URL.
7. Backlinks — overview, referring domains, anchor text, or competitor backlink overlap for a target domain or URL.
8. Site audit — list projects, then read audit status, snapshots, issues, and affected pages for technical SEO triage.

Data model

Regional database → Domain / URL / Phrase → Metrics rows. Keyword and domain reports anchor on a two-letter market code (us, gb, de, …). Backlink reports anchor on a target (root domain, subdomain, or page URL). Site audit data hangs off project ids in Semrush Projects.

Operational order

1. Pick the regional database that matches the user's market before keyword or domain pulls; use worldwide overview when the question spans all markets.
2. Start with a small row limit on unfamiliar domains or long keyword lists — each report consumes API units by type and volume.
3. Run competitor lists before deep organic or paid keyword exports when the goal is gap analysis.
4. For off-page questions, pull backlinks overview first, then referring domains or anchors as needed.
5. For technical SEO, list projects to resolve a project id, then read audit info and the latest snapshot before drilling into issue lists.
6. Do not guess project ids, snapshot ids, or issue ids — take them from list responses.

Domain and keyword research

Use domain overview for executive snapshots (traffic estimate, keyword count, authority). Organic and paid keyword reports expose what a site ranks for; competitor reports show who shares those keywords. Phrase-centric workflows combine keyword overview, related keywords, and difficulty. Question-keyword research uses phrase-question reports for content briefs.

Backlink research

Frame backlink questions around a target and whether the scope is the whole root domain, a subdomain, or one URL. Overview summarizes link volume and authority signals; referring domains and anchor reports support link-quality and over-optimization audits; competitor overlap compares off-page profiles between domains.

Site audit triage

Projects with Site Audit enabled expose crawl status, historical snapshots, categorized issues, and pages affected by each issue. Launching a new crawl or enabling Site Audit on a project are write actions that consume API units and require confirm_write after user approval. Analytics and backlink data are read-only — rankings and link graphs cannot be edited through Semrush.

Gotchas

• API units — wide exports and repeated probes burn credits fast; check balance early and cap row limits on first passes.
• Regional databases — mixing markets without setting database skews keyword and traffic numbers.
• Raw vs parsed — some legacy domain reports return semicolon-separated text; typed reads parse when possible and still include raw payload for audit.
• Site Audit tier — project management and crawl launch need a Semrush plan with Projects access; expect permission errors on lower tiers.
• Writes need confirmation — enabling audit or launching a crawl mutates project state; expect confirm_write on those actions.