AEO Scanner — MCP Server
AI search visibility audit for any website. Three scores, one scan.
What it does
- AEO Score (0-100): How well AI search engines (ChatGPT, Perplexity, Google AI Overviews) can find, read, and cite your content
- GEO Score (0-100): How likely AI is to cite your site as a source — citation readiness for AI-generated answers
- Agent Readiness (0-100): How easily AI agents can understand, interact with, and transact on your site
- Mention Readiness (0-100): Predicts how likely AI will mention your brand, based on 12 research-backed signals
- AI Identity Card: How AI currently perceives your brand — gaps, citable claims, detected competitors
- Business Profile: Auto-detected business type (commerce/saas/media/general) with which scores matter most
Quick start
Remote server (recommended) — zero install, just add the URL:
# Claude Code
claude mcp add --transport http aeo-scanner https://aeo-mcp.convrgent.ai/mcp
# Claude Desktop / Claude.ai
# Settings → Connectors → Add custom connector → https://aeo-mcp.convrgent.ai/mcp
# Cursor — add to .cursor/mcp.json
{ "aeo-scanner": { "url": "https://aeo-mcp.convrgent.ai/mcp" } }
Then ask your AI assistant: "Scan example.com for AI visibility"
Alternative: local install via PyPI (stdio transport):
claude mcp add aeo-scanner -- uvx aeo-scanner
Tools
| Tool | What it does | Price |
|---|---|---|
scan_site | Quick triple-score scan + AI Identity Card + mention readiness + business profile + top issues | Free |
audit_site | Full 58+ check breakdown across 12 categories (4 AEO + 4 GEO + 4 Agent) | $1.00 |
compare_sites | Two-site gap analysis + category winners + overtake fix code | $3.00 |
fix_site | Generated fix code with two-tier score projections — apply directly with Claude Code | $5.00 |
Free tier
scan_site works without any authentication. No API key, no wallet, no setup.
Rate limits: 20 scans/hour per IP, 5 per URL per day.
Paid tools
audit_site, compare_sites, and fix_site require an API key:
- Get your key at scan.convrgent.ai
- Set
AEO_API_KEYin your MCP config
Or pay per call with USDC via x402 protocol (Base network).
Workflow
The included optimize_site prompt guides the full workflow:
- Scan — get baseline scores + mention readiness + business profile (free)
- Compare — benchmark against detected competitors ($3)
- Audit — see detailed breakdown by category ($1)
- Fix — get working code with quick-win and full-ceiling projections ($5)
- Rescan — verify improvement across all scores (free)
Scoring
58+ checks across 12 categories. See the built-in aeo://reference/scoring-methodology resource for full details.
AEO categories: Structured Data (30%), Meta & Technical (20%), AI Accessibility (25%), Content Quality (25%)
GEO categories: Brand Narrative Clarity (25%), Citation Readiness (25%), Authority Signals (25%), Entity Definition (25%)
Agent Readiness categories: Machine Identity (30%), API Discoverability (25%), Structured Actions (25%), Programmatic Access (20%)
Grades: A (90+), B (75-89), C (60-74), D (40-59), F (0-39)
Research basis
GEO checks are backed by peer-reviewed research: Princeton KDD 2024 (+115% visibility from citing sources), Growth Memo (44.2% citations from first 30% of content), Cornell (+28% from quantitative claims), and more. See scoring methodology for full citations.
Built by Convrgent — AI visibility tools for agents.
