MCP Documentation
Model Context Protocol (MCP 2026-07-28) lets AI agents run SEO audits, get opportunities, and generate fixes via a standardized interface. Built with the official MCP TypeScript SDK.
What is MCP?
Model Context Protocol (MCP) is a standard for connecting AI agents to external tools. AuditMe implements MCP 2026-07-28, so any compatible client (Claude Desktop, Cursor, Windsurf, etc.) can call these tools:
- • SEO audits on any public URL (159 checks)
- • Prioritized issue lists ranked by impact
- • Deterministic fix suggestions (template-based, no LLM)
- • Audit methodology and SEO concepts via resources
- • Pre-built prompt templates for common workflows
- • Competitor comparisons
Auth
Public: get_site_info, get_audit_methodology, get_seo_concept — no key needed. Write: seo_audit, workflow_execute, etc. — require x-api-key header. Keys are your account API keys (or operator keys); a revoked or disabled key stops working within 60 seconds.
Setup
Connect any MCP-compatible client to AuditMe:
{
"mcpServers": {
"auditme": {
"url": "https://www.auditme.dev/api/mcp",
"transport": "streamable-http",
"headers": {
"x-api-key": "your-api-key-here"
}
}
}
}curl https://www.auditme.dev/.well-known/mcp.json
curl -X POST https://www.auditme.dev/api/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-H "MCP-Protocol-Version: 2026-07-28" \
-H "Mcp-Method: tools/list" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{"_meta":{"io.modelcontextprotocol/protocolVersion":"2026-07-28","io.modelcontextprotocol/clientCapabilities":{}}}}'curl -X POST https://www.auditme.dev/api/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-H "MCP-Protocol-Version: 2026-07-28" \
-H "Mcp-Method: tools/call" \
-H "Mcp-Name: get_seo_concept" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_seo_concept","arguments":{"concept":"canonical-url"},"_meta":{"io.modelcontextprotocol/protocolVersion":"2026-07-28","io.modelcontextprotocol/clientCapabilities":{}}}}'curl -X POST https://www.auditme.dev/api/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-H "MCP-Protocol-Version: 2026-07-28" \
-H "Mcp-Method: tools/call" \
-H "Mcp-Name: seo_audit" \
-H "x-api-key: your-api-key" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"seo_audit","arguments":{"url":"https://example.com"},"_meta":{"io.modelcontextprotocol/protocolVersion":"2026-07-28","io.modelcontextprotocol/clientCapabilities":{}}}}'Available Tools (14)
get_site_infoPublicGet basic site information: title, description, tech stack, crawl status.
- • url (string, required)
Site metadata object with title, description, tech stack, and crawl status.
get_audit_methodologyPublicGet the full audit methodology: 159 checks across 16 categories, severity logic, data sources, and limitations.
- • category (string, optional)
Methodology details with weights, severity definitions, and known limitations.
get_seo_conceptPublicGet explanation of an SEO concept with definition, why it matters, how AuditMe checks it, and best practices.
- • concept (enum, required): canonical-url | meta-description | structured-data | core-web-vitals | toxic-backlink | robots-txt | xml-sitemap
Concept explanation with definition, importance, audit check details, and best practices.
seo_auditAuth requiredRun full unified audit (159 checks across 16 categories) on a URL. Returns score, grade, issues by severity, and top issues.
- • url (string, required)
Score (0-100), grade (A-F), issue counts by severity, top 5 issues with details.
get_opportunitiesAuth requiredPrioritize audit issues by impact, effort, and business relevance. Uses opportunity-weighted scoring.
- • url (string, required)
Sorted issue list with priority scores, categories, and fix recommendations.
generate_fixesAuth requiredGenerate deterministic fix suggestions for top audit issues. No AI hallucination — all fixes are data-driven.
- • url (string, required)
- • topN (number, optional, default 5, max 20)
Fix suggestions with code examples and priority.
workflow_executeAuth requiredExecute a full workflow: audit, analyze, fix, monitor, or competitor comparison.
- • intent (enum, required): audit | analyze | fix | monitor | competitor
- • url (string, required)
- • competitorUrl (string, optional)
- • topN (number, optional)
Workflow run ID, status, step progress, and output.
workflow_statusAuth requiredCheck status and results of a workflow run.
- • runId (string, required)
Run status, step details with errors, and output.
approve_fixesAuth requiredApprove or reject pending fix changes from a workflow.
- • runId (string, required)
- • approvalId (string, required)
- • decision (enum, required): approved | rejected
Updated run status and output.
detect_regressionsAuth requiredCompare two audit results and detect regressions.
- • baselineUrl (string, required)
- • currentUrl (string, required)
List of regressions with severity and descriptions.
compare_competitorsAuth requiredCompare your site against a competitor side by side.
- • url (string, required)
- • competitorUrl (string, required)
Comparison with score differences and recommendations.
monitor_siteAuth requiredGet monitoring overview for a site: alert rules, recent alerts, and severity breakdown.
- • url (string, required)
Alert rules, recent alerts list, severity counts, and monitoring status.
track_keywordsAuth requiredAdd, remove, or list tracked keywords for a site. Monitor ranking changes over time.
- • action (enum, required): add | remove | list
- • keyword (string, required for add/remove)
- • targetUrl (string, required for add)
For add: confirmation with keyword ID. For remove: confirmation. For list: array of tracked keywords with last-known positions.
run_experimentAuth requiredCreate or list SEO experiments with hypothesis, baseline, and metric tracking.
- • action (enum, required): create | list
- • title (string, required for create)
- • hypothesis (string, required for create)
- • url (string, required for create)
- • experimentType (string, optional)
- • metricTarget (string, optional)
For create: experiment ID and baseline snapshot. For list: array of experiments with status.
MCP Resources
Resources provide static knowledge that AI agents can read for context. Access them via auditme:// URIs.
auditme://methodologyauditme://seo-conceptsauditme://checksauditme://severity-levelsauditme://data-sourcesMCP Prompts
Pre-built prompts guide AI agents through common SEO workflows. Use them as starting points.
audit_websiteRun a full SEO audit on a URL and report findings with recommendations.
explain_seo_issueExplain an SEO issue found during audit, why it matters, and how to fix it.
prioritize_seo_fixesRank audit issues by impact and effort, then propose a fix schedule.
analyze_ai_visibilityEvaluate how well a site works for AI search engines (ChatGPT, Perplexity, Google AI Overviews).
Security
Example: Full Audit via MCP
Send the body below as POST /api/mcp with headers Content-Type: application/json, MCP-Protocol-Version: 2026-07-28, Mcp-Method: tools/call, Mcp-Name: seo_audit and (for write tools) x-api-key. The response comes back as a single JSON-RPC message with Content-Type: application/json.
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "seo_audit",
"arguments": {
"url": "https://example.com"
},
"_meta": {
"io.modelcontextprotocol/protocolVersion": "2026-07-28",
"io.modelcontextprotocol/clientCapabilities": {}
}
}
}{
"jsonrpc": "2.0",
"id": 1,
"result": {
"content": [{
"type": "text",
"text": "{\n \"url\": \"https://example.com\",\n \"score\": 82,\n \"grade\": \"B\",\n \"passed\": 106,\n \"total\": 159,\n \"issues\": { ... },\n \"topIssues\": [ ... ]\n}"
}],
"structuredContent": {
"url": "https://example.com",
"score": 82,
"grade": "B",
"passed": 106,
"total": 159,
"issues": { "total": 25, "critical": 1, "high": 3, "medium": 8, "low": 13 },
"topIssues": ["..."]
}
}
}