What is MCP for AI visibility versus MCP for SEO?

MCP for SEO is SellOnLLM’s connector at https://www.sellonllm.com/api/mcp—it reads your GA4 and Google Search Console data. MCP for AI visibility is this connector at https://www.sellonllm.com/api/mcp-ai-visibility—it focuses on AEO site crawls and optional Perplexity citation tools. Use both only if your Claude plan allows multiple custom connectors.

Where do I save my Perplexity API key?

Sign in on sellonllm.com with the same Google account you use for the AI visibility MCP, open this page, and save your key in the Perplexity section. Keys are encrypted at rest. analyze_website_aeo does not require Perplexity.

What is the AI visibility MCP server URL for Claude?

Use https://www.sellonllm.com/api/mcp-ai-visibility when adding a custom connector. OAuth discovery uses the same authorization server as other SellOnLLM MCP connectors, with resource-specific metadata.

Model Context Protocol · Claude · AEO

AI visibility & AEO MCP for Claude

This is SellOnLLM’s MCP for AI visibility and answer-engine optimization (AEO): add it free in Claude for crawl-based scorecards and optional Perplexity citation checks. For GA4 + Google Search Console workflows, use our separate MCP for SEO instead. Practical write-ups live in the SellOnLLM blog.

Not the GA connector — use GA + GSC MCP for properties, queries, CTR, and traffic tools.

Connect in Claude

Paste only the AI visibility server URL below. OAuth discovery is the same family as our other MCP, but tokens are scoped to this resource.

Open claude.ai → Settings → Connectors.
Add custom connector (this is separate from any GA+GSC connector you may already have).
Paste the URL below → Connect → sign in with Google and approve when prompted.

Claude Free allows only one custom connector per account. If you need both GA+GSC and AI visibility in Claude, use a plan that supports multiple connectors, or use the web chat for analytics while reserving your connector slot for AI visibility.

Perplexity API key not set

check_ai_visibility, discover_ranking_prompts, get_visibility_report, and compare_competitor_visibility call Perplexity with your key. analyze_website_aeo only crawls HTML and does not use Perplexity.

Tools on this connector

analyze_website_aeo — Crawl public pages and score AEO signals (schema, FAQ, depth, trust).
check_ai_visibility — One query vs one URL in Perplexity Sonar citations.
discover_ranking_prompts — Generate prompts from the site, then test visibility (rate-limited).
get_visibility_report — Combined report with optional custom queries and auto-generated prompts.
compare_competitor_visibility — Same prompts across your URL and competitors (hosted caps apply).

Copy/paste prompts

After connecting, Claude can also use / → aeo_site_audit, ai_visibility_pulse, or research_first_visibility (recommended for realistic buyer prompts). Ask Claude to render a dashboard-style artifact when supported, with markdown fallback for clients that do not support artifacts.

AEO site audit

Crawl + scorecard

Use analyze_website_aeo on my primary marketing site (I will paste the URL next). Render the result as a dashboard-style artifact when supported by Claude with KPI cards, a signal breakdown section, a pages-crawled table, and the top 5 fixes in priority order. If artifacts are unavailable, fall back to markdown tables and bullets.

AI visibility pulse

Perplexity checks for a few prompts

I will give you my site URL and 2–3 customer questions we care about. Use check_ai_visibility for each question and render a compact dashboard artifact when supported: KPI cards for visible/not visible, a table of prompts vs citation position, and concise recommendations. If artifacts are unavailable, fall back to markdown tables and bullets.

Research-first visibility (recommended)

If browsing is available: research business + competitors, then test realistic prompts

Do this in three phases:

Phase 1 — Research (if browsing is available):
- Look up my company using the URL I provide next.
- Infer: what we sell (offerings), target audience/persona, geographic focus, and 2–5 close competitors.
- Identify 8–12 realistic buyer-intent prompts people would type into ChatGPT/Claude/Perplexity (pricing, alternatives, reviews, vs competitor, best-for-persona, local intent).

Phase 2 — MCP testing:
- Use discover_ranking_prompts on my URL with:
  - offerings (3–6 items)
  - audience (1 short phrase)
  - geo (if relevant)
  - competitors (2–5)
  - seed_prompts (your best 6–8 buyer-intent prompts)
  - max_prompts 8
- Then run check_ai_visibility on the top 3 most important prompts.

Phase 3 — Output:
- Render a dashboard-style artifact when supported by Claude.
- Use KPI cards for AEO score / visibility rate / prompts tested / wins.
- Include a detailed table: prompt | visible? | who is cited instead | 1 specific fix.
- End with a prioritized 14-day action plan (impact × effort).
- If artifacts are unavailable, fall back to markdown tables and bullets.

I will paste the website URL next.

Frequently asked questions

Is this the same MCP as Google Analytics and Search Console?

No. Analytics tools live on https://www.sellonllm.com/api/mcp. This page is for https://www.sellonllm.com/api/mcp-ai-visibility only.

Do I need a Perplexity key?

Only for Perplexity-backed tools. analyze_website_aeo works without it. Get a key at perplexity.ai/settings/api (must start with pplx-).

Can I use both MCPs in Claude?

Yes if your Claude plan allows multiple custom connectors: add one connector URL for analytics and a second for AI visibility. Each issues its own MCP access token tied to the correct resource.