MCP Integration Examples

LLMSE provides a public Model Context Protocol (MCP) server that allows AI assistants like Claude to classify websites and match advertisers. Below are examples and prompts you can try.

Quick Start

Add LLMSE to your AI assistant using the command line:

Claude Code

claude mcp add --transport http llmc-public https://llmse.ai/mcp

This adds the server to your local scope. Use --scope project to share with your team via .mcp.json.

Gemini CLI

gemini mcp add llmc-public https://llmse.ai/mcp --transport http

By default, the server is added to ~/.gemini/settings.json. Use --scope project for project-specific config.

OpenAI Codex

Edit ~/.codex/config.toml and add:

[mcp_servers.llmc-public]
url = "https://llmse.ai/mcp"

This configuration is shared between Codex CLI and the VSCode extension.

Available Tools

audit — Comprehensive website audit combining classification, SEO analysis, and advertiser matching in one efficient call
classify_url — Classify any website into category, subcategory, language, demographics, and sentiment
select_advertiser — Match advertising networks to website demographics based on 98 real ad networks
analyze_seo — Analyze any website for SEO issues with scoring, grading, and recommendations

audit Tool

Performs a comprehensive website audit that combines classification, SEO analysis, and advertiser matching in a single efficient call. This is the recommended tool when you need a complete analysis since it fetches the page only once.

Parameters

url The website URL to audit (required)

Example Prompts

"Run a full audit on https://techcrunch.com using LLMSE"
"Audit https://espn.com and show me classification, SEO, and best advertisers"
"Give me a complete analysis of https://nytimes.com"
"Use the audit tool on https://mywebsite.com to check SEO and find matching ad networks"
"Comprehensive website report for https://bloomberg.com"

Response Structure

classification Category, subcategory, language, sentiment, age, and gender demographics
seo Score (0-100), grade (A-F), issues by severity, meta info, and recommendations
advertisers Top 3 matched advertising networks with scores based on demographics
cached Whether the result was from cache

Why Use Audit?

Efficiency Fetches the page once instead of three separate requests
Complete Picture Get classification, SEO issues, and ad network recommendations together
Consistent Data All analyses based on the same page snapshot

classify_url Tool

Classifies a website URL and returns detailed information about its category, language, target audience, and sentiment.

Parameters

url The website URL to classify (required)

Example Prompts

"Classify https://techcrunch.com using LLMSE"
"What category is https://espn.com?"
"Analyze https://wikipedia.org with the llmc-public MCP"
"Use classify_url on https://nytimes.com and tell me about its audience"
"What language and sentiment does https://lemonde.fr have?"

Response Fields

category Main category (e.g., "Technology", "Sports", "News and Media")
subcategory Specific subcategory within the main category
language Primary content language (e.g., "English", "Spanish")
age Target age group (e.g., "18-24", "25-34", "30-50")
gender Target gender ("male", "female", or "all")
sentiment Content sentiment ("Good", "Neutral", or "Bad")
cached Whether the result was from cache

select_advertiser Tool

Matches advertising networks to website demographics. Can work with a URL (fetches classification) or direct demographic parameters.

Parameters

url URL to match advertisers for (uses cached classification)
category Target category (e.g., "Sports", "Technology")
subcategory Target subcategory
age Target age group (e.g., "18-24", "25-34")
gender Target gender ("male", "female", or "all")
sentiment Content sentiment ("Good", "Neutral", or "Bad")
limit Number of advertisers to return (1-10, default 3)

Example Prompts - By URL

"Select advertisers for https://espn.com using llmc-public"
"Find the best ad networks for https://techcrunch.com"
"Which advertisers match https://vogue.com? Show me 5 options"
"Use select_advertiser for https://wikipedia.org/wiki/Python"

Example Prompts - By Demographics

"Find advertisers targeting Sports category, male audience, ages 18-24"
"Which ad networks target Beauty category with female audience?"
"Select 5 advertisers for Technology category with Good sentiment"
"Match advertisers for Finance category targeting 35-44 age group"

Scoring Algorithm

Advertisers are scored based on demographic matches:

+10 pts Category match
+5 pts Age group match
+3 pts Gender match
+2 pts Sentiment match (brand safety)
tiebreaker Higher CPM bid

analyze_seo Tool

Analyzes a website URL for SEO issues and provides an overall score, letter grade, and actionable recommendations.

Parameters

url The website URL to analyze (required)

Example Prompts

"Analyze SEO for https://example.com using LLMSE"
"What SEO issues does https://techcrunch.com have?"
"Run an SEO audit on https://mywebsite.com and give me recommendations"
"Check the SEO score of https://blog.example.com"
"Use analyze_seo to find meta tag issues on https://store.example.com"

Response Fields

score Overall SEO score from 0-100
grade Letter grade (A, B, C, D, or F)
issues.critical Critical issues that severely impact SEO (-15 points each)
issues.warnings Warnings that should be addressed (-5 points each)
issues.info Informational items for consideration (-1 point each)
meta Extracted metadata (title, description, headings, images, links, etc.)
recommendations Prioritized list of improvements sorted by severity
cached Whether result was from cache

SEO Checks Performed

Title Tag Presence, length (30-60 chars recommended)
Meta Description Presence, length (120-160 chars recommended)
Headings H1 presence/count, heading hierarchy
Technical Viewport, canonical, charset, language attribute
Robots Noindex/nofollow detection
Images Alt attribute presence and quality
Social Open Graph tags, Twitter Cards
Structured Data JSON-LD schema presence and types
Content Word count (thin content detection)
Links Internal vs external link analysis

Advertising Networks

LLMSE includes 98 real advertising networks across 11 verticals:

DSPs (10) Google Ads, DV360, Amazon DSP, The Trade Desk, Microsoft, Yahoo, StackAdapt, Viant, Simpli.fi, Quantcast
Social (7) Meta Ads, LinkedIn, TikTok, Snapchat, Pinterest, Reddit, X
CTV/OTT (14) Roku, Samsung, LG, VIZIO, Hulu, Peacock, Paramount, Disney+, Netflix, Tubi, Pluto TV, MNTN, Tatari, Vibe
Retail (12) Walmart Connect, Target Roundel, Instacart, Kroger, Albertsons, CVS, Best Buy, eBay, Mercado Libre, Alimama, Home Depot, Sam's Club
Gaming (11) AppLovin, Unity, Twitch, Anzu, Overwolf, AdMob, InMobi, Digital Turbine, Liftoff, ironSource, Moloco
Audio (7) Spotify, iHeartMedia, SiriusXM/Pandora, Acast, Amazon Audio, AudioGO, Podbean
Native (9) Taboola, Outbrain, MGID, Revcontent, TripleLift, Sharethrough, Criteo, Yahoo Gemini, Nativo
B2B (7) Demandbase, 6sense, RollWorks, Terminus, Madison Logic, ZoomInfo, Metadata.io
Specialized (9) DeepIntent, PulsePoint, Sojern, Adara, Dianomi, Healthgrades, Doximity, Cardlytics, TripAdvisor
SSPs (6) Magnite, PubMatic, OpenX, Index Exchange, Xandr, Sovrn
DOOH (6) Vistar Media, Broadsign, Clear Channel, OUTFRONT Media, JCDecaux, Lamar

Advanced Use Cases

Competitive Analysis

"Classify these competitor sites: https://cnn.com, https://foxnews.com, https://bbc.com - compare their categories and audiences"

Ad Campaign Planning

"I want to advertise to males 18-24 interested in Gaming. Which ad networks should I use?"
"Find the best CTV advertisers for Entertainment content with Good sentiment"

Content Analysis

"What's the sentiment of https://reddit.com/r/technology?"
"Classify this URL and explain who the target audience is: https://arstechnica.com"

Bulk Operations

"Classify these 5 URLs and create a comparison table: [list of URLs]"
"For each of these news sites, find matching advertisers and rank by CPM"

Technical Details

Protocol Model Context Protocol (MCP) 2024-11-05
Endpoint https://llmse.ai/mcp
Transport Streamable HTTP with SSE
Categories 58 main categories with hundreds of subcategories
Caching Results are cached for fast repeated lookups

Back to Documentation

See the About page for general information about LLMSE, or visit Query Examples for web interface usage.