AudaeoMCP Integration Guide

Connect AI Agents to Your Audit Data

The Audaeo MCP server lets AI assistants like Claude access your audit results directly. Ask questions, get recommendations, analyze page content, and generate action plans — all powered by your real CLEAR Framework data.

What is MCP?

Model Context Protocol (MCP) is an open standard that lets AI assistants use tools hosted on external servers. Think of it as giving Claude a set of specialized functions it can call to pull data from your AEO audits in real-time.

Instead of copy-pasting audit results into a chat, Claude can directly query your audit database, read page content, compare scores, and generate specific recommendations — all within a natural conversation.

Setup in 2 Minutes

1

Get Your API Key

Go to your Settings page to request an API key for MCP access, or contact your account administrator.

2

Add the Configuration

Choose your AI tool below and add the configuration. Replace YOUR_API_KEY with your actual key.

Open ~/Library/Application Support/Claude/claude_desktop_config.json and add:

{
  "mcpServers": {
    "aeo-audit": {
      "type": "url",
      "url": "https://app.audaeo.com/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}
3

Start Using It

Restart your AI tool and start asking questions about your audits. Claude will automatically call the right tools to get the data it needs.

Available Tools

16 tools across audit analysis, SEO intelligence, and CRM-enriched insights.

Audit & CLEAR

8 tools

Core audit data, page analysis, CLEAR Framework scores, and recommendations

get_clear_framework

Returns the full CLEAR Framework definition with scoring rubrics. Use this to understand what the audit measures.

Try: “Explain the CLEAR framework and how scores are calculated”

list_audits

Lists all audits with domain, grade, score, and status.

Try: “What audits have been run?”

get_auditdomain (required)

Full audit details for a domain — scores, impression, strengths, weaknesses, recommendations.

Try: “Show me the audit results for impulsecreative.com”

get_page_analysisdomain (required), url (optional)

CLEAR scores and Human/AI perspectives for each page. Omit URL to get all pages.

Try: “How does the homepage score on Evidence & Emotion?”

get_recommendationsdomain (required)

Prioritized recommendations — site-level synthesis plus per-page action items.

Try: “What are the top 5 things to fix on our site?”

get_page_contentdomain (required), url (required)

The actual markdown content from the last scrape. Useful for drafting improvements.

Try: “Show me the content on our pricing page and suggest improvements”

compare_dimensionsdomain (required)

Compare CLEAR scores across all pages — find consistent weaknesses.

Try: “Which dimension is weakest across the whole site?”

get_executive_summarydomain (required)

Grade story with page distribution, dimension weaknesses, query coverage vs score disconnect, and content brief counts.

Try: “Give me the executive summary for our audit”

SEO & AI Search

4 tools

Keywords, query fan-out simulation, content gaps, and keyword recommendations

get_keywordsdomain (required), funnel_stage (optional), min_volume (optional)

SEO keyword data — what the site ranks for, positions, search volume, difficulty, and funnel stage.

Try: “What keywords do we rank for with over 1000 monthly searches?”

get_query_fanoutdomain (required), gaps_only (optional)

AI search query fan-out analysis — how AI engines decompose buyer questions into sub-queries and whether your site has content for each.

Try: “Show me all the buyer questions where we have no content coverage”

get_content_gapsdomain (required), funnel_stage (optional), include_page_fanouts (optional)

AI search visibility gaps grouped by funnel stage. Includes per-page fan-out data showing which pages cover which sub-queries.

Try: “What BOFU content gaps do we have?”

get_keyword_recommendationsdomain (required), source (optional), min_volume (optional)

Keywords the site SHOULD be targeting but isn't — sourced from page content analysis and fan-out gaps.

Try: “What new keywords should we target based on our content gaps?”

CRM Insights

3 tools

HubSpot CRM data integration — deals, tickets, contacts, and GTM recommendations

get_crm_insightsdomain (required)

Raw CRM data from HubSpot — deals (stage/industry/amount), tickets (categories/priorities), contacts (lifecycle stages/traffic sources), and companies. Scoped to the audited domain.

Try: “What CRM data do we have for this domain? Show me the deal breakdown.”

get_crm_recommendationsdomain (required), severity (optional: critical/high/medium/info)

AI-synthesized GTM recommendations connecting CRM data to AEO gaps — with severity, evidence, and specific data points from deals, tickets, and contacts.

Try: “What are the critical CRM-backed recommendations for our site?”

get_crm_summarydomain (required)

Combined GTM + AEO snapshot — audit score, CRM data counts, and top 5 recommendations in one view. Designed for quick business impact assessment.

Try: “Give me the GTM + AEO picture for our domain”

Example Conversations

"Pull up the audit for our site and give me the 3 highest-impact fixes"

Claude calls get_audit and get_recommendations, then prioritizes by score impact

"Look at our homepage content and write the missing Organization schema markup"

Claude calls get_page_content for the homepage, analyzes it, and generates JSON-LD

"Which of our pages has the worst AI Agent score and why?"

Claude calls compare_dimensions, finds the lowest scorer, then calls get_page_analysis for details

"Show me our content gaps for bottom-of-funnel buyer questions"

Claude calls get_content_gaps with funnel_stage=BOFU to find where AI search can't cite you at the decision stage

"What industries are our closed-won deals in, and do we have content for them?"

Claude calls get_crm_insights for deal data, then cross-references with get_query_fanout to find industry content gaps

"Give me the full GTM + AEO picture and build a 90-day content plan"

Claude calls get_crm_summary for the snapshot, then get_crm_recommendations and get_content_gaps to build a prioritized plan

For AI Agent Developers

The MCP server supports automatic discovery. Point your MCP client at:

https://app.audaeo.com/.well-known/mcp.json

This returns the server manifest with all available tools, authentication requirements, and connection details. The server uses the Streamable HTTP transport protocol.