SEO MCP Prompt Library
Use task-specific SEO MCP prompts for crawls, links, images, rankings, AI visibility, and safe project workflows.
Good SEO MCP prompts define the outcome, evidence scope, safety boundary, and output format. They do not need to name every tool; the server already instructs agents to discover projects and crawls in the correct order.
Technical SEO audit
Audit the latest completed crawl for [PROJECT].
Start with the overview, then verify the highest-impact issues with records.
For every recommendation include: evidence, affected URL pattern, impact,
owner, smallest fix, and post-release validation.
Read-only analysis. Do not start a crawl or change settings.HTTP status audit
Find all non-2xx HTML pages in the latest completed crawl.
Group by status code and URL pattern. Separate final URLs from redirects.
Prioritize pages with strong internal-link exposure and return up to 20 examples.Internal link repair
Find failed internal-link destinations.
Group repeated targets and retrieve their source pages.
Identify navigation or template-wide patterns before isolated content links.
Draft a repair plan with acceptance criteria.Image SEO queue
Find failed images, unusually large assets, and missing-alt candidates.
Group shared assets by source-page count.
Separate technical facts from editorial decisions about alt text.Crawl comparison
Compare crawl [BEFORE_UID] with [AFTER_UID].
Return improvements, regressions, expected changes, and uncertain changes.
Verify major deltas with page, link, or image records.
Include both crawl UIDs and timestamps in the report.Core Web Vitals review
Review Core Web Vitals for [PROJECT].
Group URLs by template. Identify weak LCP, INP, and CLS patterns.
Suggest the next diagnostic check, but do not claim an unproven root cause.Multi-project prioritization
Compare the projects I can access.
Rank where technical SEO attention is most likely to help.
Explain the evidence, confidence, and missing data behind the ranking.Tracked keyword change
For [PROJECT], compare the latest two stored Rank Tracker analyses for
[KEYWORD] in [LOCATION] on [DEVICE]. Show dates, positions, and landing URLs.
Explain any missing analysis or changed configuration. Do not add keywords.AI Visibility evidence
For [PROJECT], list monitored prompts that mention [BRAND] in the selected
analysis period. Open representative answers and sources. Separate a brand
mention from a cited source and note the model and date. Do not add prompts.Ticket creation
Turn the verified findings into engineering tickets.
Group records by shared root cause, not one ticket per URL.
Include evidence, representative URLs, proposed change, acceptance criteria,
risk, owner, and the exact SEO MCP prompt to validate the release.Safe crawl request
Show the exact project, settings, and expected effect of starting a crawl.
Do not start it. Wait for my explicit confirmation.Prompt pattern
Reuse this structure:
Goal: [decision or outcome]
Scope: [project, crawl, page group, time window]
Evidence: [records and fields required]
Output: [table, grouped findings, tickets, summary]
Safety: [read-only or exact approved action]
Limits: [sample size, uncertainty, exclusions]For step-by-step interpretation, use the workflow hub. For tool fields and account changes, use the tools reference.
Related documentation
Continue exploring SEO MCP
SEO MCP Guides
Learn how to use SEO MCP responsibly, write verifiable prompts, audit content for search and AI answers, and choose MCP or an API.
Responsible SEO Automation
Build safe SEO MCP automations with read-only defaults, explicit confirmation, bounded queries, reproducible evidence, and human review.