MCP server: AEO / AI-search audit of any URL (direct answers, JSON-LD, llms.txt, E-E-A-T, AI citations). By Les Créavores.
- ✓Open-source license (MIT)
- ✓Actively maintained (<30d)
- ✓Clear description
- ✓Topics declared
- ✓Documented (README)
claude mcp add creavores-aeo-mcp -- npx -y creavores-aeo-mcp{
"mcpServers": {
"creavores-aeo-mcp": {
"command": "npx",
"args": ["-y", "creavores-aeo-mcp"]
}
}
}MCP Servers overview
# creavores-aeo-mcp
An MCP (Model Context Protocol) server that audits any URL for **AEO (Answer Engine Optimization)** and AI-search readiness — a practical diagnostic tool, not a gadget. It checks the same things ChatGPT, Perplexity, Google AI Overviews, and agentic browsers actually look for: a citable direct answer, valid JSON-LD, `llms.txt` compliance, E-E-A-T trust signals, and whether AI crawlers/agents can read and navigate the page at all.
Built by [Les Créavores](https://lescreavores.fr), a web & SEO/AEO agency. The scoring logic is compatible with (and extends) the public [`@macalc/aeo-score`](https://www.npmjs.com/package/@macalc/aeo-score) package and the [free AEO audit tool](https://lescreavores.fr/outils/audit-aeo-gratuit/) — scores you get here match what you'd see there.
## Installation
No install step needed — run it directly with `npx`:
```bash
npx creavores-aeo-mcp
```
### Claude Desktop
Add to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"creavores-aeo": {
"command": "npx",
"args": ["-y", "creavores-aeo-mcp"]
}
}
}
```
### Claude Code
```bash
claude mcp add creavores-aeo -- npx -y creavores-aeo-mcp
```
Or add it to `.mcp.json` at the root of your project:
```json
{
"mcpServers": {
"creavores-aeo": {
"command": "npx",
"args": ["-y", "creavores-aeo-mcp"]
}
}
}
```
No API key, no configuration, no environment variables required. The server fetches pages directly with a self-identifying User-Agent (`creavores-aeo-mcp/1.0 (+https://lescreavores.fr)`), a 15-second timeout, and never throws — every tool returns clean JSON (including `{ "error": "..." }` on failure) so it's safe to call from any agent loop.
## Tools
### 1. `audit_aeo`
Full AEO audit: crawlability (30 pts) + content structure for AI extraction (35 pts) + E-E-A-T trust (35 pts), plus `llms.txt` and AI-bot access checks. Returns a 0-100 score, a verdict, and prioritized recommendations.
```json
{ "url": "https://example.com/pricing" }
```
→ `{ score: 78, verdict: "good", subScores: {...}, topRecommendations: [...], issues: [...] }`
### 2. `check_direct_answer`
Detects a 40-60 word direct-answer paragraph right under the H1 — the strongest single citability signal for AI engines.
```json
{ "url": "https://example.com/pricing" }
```
### 3. `extract_structured_data`
Extracts and validates every JSON-LD block: types found, parse errors, common schema.org mistakes (missing `acceptedAnswer`, missing `datePublished`, etc.), and confirms the data ships server-side (not injected client-side via `useEffect`, which many AI crawlers never execute).
```json
{ "url": "https://example.com/pricing" }
```
### 4. `check_llms_txt`
Checks existence and [llmstxt.org](https://llmstxt.org/) spec compliance: H1 title, blockquote summary, `## Section` headings, and Markdown link list items.
```json
{ "domain": "example.com" }
```
### 5. `check_eeat_signals`
Byline/author, publish/modified dates, source citations, About/Methodology page links, legal-notice links, Organization schema.
```json
{ "url": "https://example.com/blog/article" }
```
### 6. `check_agentic_readiness`
**Differentiator tool**, aligned with Lighthouse's new "Agentic Browsing" audit category: is this page readable and navigable by an AI *agent* (not just a search crawler)? Checks `llms.txt`, AI-crawler access in `robots.txt`, heading hierarchy, image alt-text coverage, ARIA landmarks, viewport meta, and canonical.
```json
{ "url": "https://example.com/app" }
```
### 7. `simulate_ai_citation`
Simulates how an AI engine would extract a citable passage: scans paragraphs, list items, table rows, and FAQ answers, scores each for self-containedness, and returns the single passage most likely to be quoted — with an explanation of why (or why not).
```json
{ "url": "https://example.com/blog/article" }
```
### 8. `compare_aeo`
Compares 2-3 URLs' AEO scores side by side — use it to benchmark yourself against a competitor.
```json
{ "urls": ["https://example.com", "https://competitor.com"] }
```
### 9. `get_prix_site`
Live French website prices from our open dataset (prices publicly displayed by agencies and freelancers): one cell per `type|unit` with `n`, distinct providers and p10 to p90 in euros. Optional `type` and `unit` filters; only publishable cells by default. Read live from [reference-prix.json](https://lescreavores.fr/data/observatoire-achat-site-internet/reference-prix.json).
```json
{ "type": "vitrine", "unit": "forfait" }
```
### 10. `get_marches_publics`
Live aggregates of French public procurement for websites (DECP, BOAMP): amounts excl. VAT (n, q1, median, q3) per category and buyer type, bids received, yearly breakdown. Optional `buyer_type` filter. Read live from `reference-prix.json` and [stats.json](https://lescreavores.fr/data/observatoire-achat-site-internet/stats.json).
```json
{ "buyer_type": "commune" }
```
### 11. `get_marronniers`
Live French editorial calendar dates (Sept 2026 to Dec 2027): world days, holidays, retail events, with sectors, source and a content angle. Optional `from` / `to` (YYYY-MM-DD), `sector` and `limit`. Read live from [marronniers-2026-2027.json](https://lescreavores.fr/data/marronniers-2026-2027.json).
```json
{ "from": "2026-12-01", "to": "2026-12-31", "sector": "commerce", "limit": 10 }
```
Every data answer carries `source` (file URL), `licence` and `generated_at` as read in the file, and a `citation` with the page to link: [Observatoire de l'achat de sites internet](https://lescreavores.fr/observatoire-achat-site-internet/) or [Calendrier des marronniers 2026-2027](https://lescreavores.fr/outils/generateur-calendrier-editorial/). If a file is unreachable, the tool returns `{ "error": "..." }`.
## About
Made by [Les Créavores](https://lescreavores.fr), a French web & SEO/AEO agency (Metz, France & Luxembourg). We build the tools we use ourselves — see the free [AEO audit tool](https://lescreavores.fr/outils/audit-aeo-gratuit/) (same scoring engine, browser version) and the [website pricing barometer](https://lescreavores.fr/prix-creation-site-internet/) (open dataset, CC-BY).
---
# creavores-aeo-mcp (Français)
Un serveur MCP (Model Context Protocol) qui audite n'importe quelle URL pour l'**AEO (Answer Engine Optimization)** et la préparation aux moteurs de recherche IA — un outil de diagnostic concret, pas un gadget. Il vérifie exactement ce que regardent ChatGPT, Perplexity, les AI Overviews de Google et les agents de navigation autonomes : une réponse directe citable, des données structurées JSON-LD valides, la conformité du fichier `llms.txt`, les signaux de confiance E-E-A-T, et la capacité des robots/agents IA à lire et naviguer réellement la page.
Développé par [Les Créavores](https://lescreavores.fr), une agence web et SEO/AEO. La logique de scoring est compatible avec le package public [`@macalc/aeo-score`](https://www.npmjs.com/package/@macalc/aeo-score) et l'outil gratuit d'[audit AEO en ligne](https://lescreavores.fr/outils/audit-aeo-gratuit/) — les scores obtenus ici correspondent à ceux affichés là-bas.
## Installation
Aucune étape d'installation nécessaire — lancez-le directement avec `npx` :
```bash
npx creavores-aeo-mcp
```
### Claude Desktop
Ajoutez ceci à votre `claude_desktop_config.json` :
```json
{
"mcpServers": {
"creavores-aeo": {
"command": "npx",
"args": ["-y", "creavores-aeo-mcp"]
}
}
}
```
### Claude Code
```bash
claude mcp add creavores-aeo -- npx -y creavores-aeo-mcp
```
Ou ajoutez-le dans le fichier `.mcp.json` à la racine de votre projet :
```json
{
"mcpServers": {
"creavores-aeo": {
"command": "npx",
"args": ["-y", "creavores-aeo-mcp"]
}
}
}
```
Aucune clé API, aucune configuration ni variable d'environnement requise. Le serveur récupère les pages directement avec un User-Agent identifiable (`creavores-aeo-mcp/1.0 (+https://lescreavores.fr)`), un délai d'attente de 15 secondes, et ne lève jamais d'exception non gérée : chaque outil renvoie un JSON propre (y compris `{ "error": "..." }` en cas d'échec), ce qui le rend sûr à appeler depuis n'importe quelle boucle d'agent.
## Les 11 outils
### 1. `audit_aeo`
Audit AEO complet : crawlabilité (30 pts) + structure du contenu pour l'extraction IA (35 pts) + confiance E-E-A-T (35 pts), plus vérification du `llms.txt` et de l'accès des robots IA. Renvoie un score sur 100, un verdict et des recommandations priorisées et actionnables.
```json
{ "url": "https://exemple.fr/tarifs" }
```
### 2. `check_direct_answer`
Détecte une réponse directe de 40 à 60 mots placée juste sous le H1 — le signal de citabilité le plus fort pour les moteurs IA.
```json
{ "url": "https://exemple.fr/tarifs" }
```
### 3. `extract_structured_data`
Extrait et valide chaque bloc JSON-LD : types trouvés, erreurs de parsing, erreurs courantes du schéma (par exemple `acceptedAnswer` ou `datePublished` manquants), et confirme que les données sont bien présentes côté serveur (et non injectées côté client via `useEffect`, que de nombreux robots IA n'exécutent jamais).
```json
{ "url": "https://exemple.fr/tarifs" }
```
### 4. `check_llms_txt`
Vérifie l'existence et la conformité à la spécification [llmstxt.org](https://llmstxt.org/) : titre H1, résumé en citation, sections `## Titre`, et liste de liens au format Markdown.
```json
{ "domain": "exemple.fr" }
```
### 5. `check_eeat_signals`
Signature/auteur, dates de publication et de mise à jour, citations de sources, liens vers une page À propos/Méthodologie, liens vers les mentions légales, schéma Organization.
```json
{ "url": "https://exemple.fr/blog/article" }
```
### 6. `check_agentic_readiness`
**Outil différenciant**, aligné sur la nouvelle catégorie d'audit Lighthouse « Agentic Browsing » : cette page est-elle lisible et navigable par un *agent* IA (pas seulement un robot d'indexation classique) ? Vérifie le `llms.txt`, l'accès des robots IA dans le `robots.tWhat people ask about creavores-aeo-mcp
What is tresor4k/creavores-aeo-mcp?
+
tresor4k/creavores-aeo-mcp is mcp servers for the Claude AI ecosystem. MCP server: AEO / AI-search audit of any URL (direct answers, JSON-LD, llms.txt, E-E-A-T, AI citations). By Les Créavores. It has 0 GitHub stars and its last recorded update is dated 2026-10-12.
How do I install creavores-aeo-mcp?
+
You can install creavores-aeo-mcp by cloning the repository (https://github.com/tresor4k/creavores-aeo-mcp) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is tresor4k/creavores-aeo-mcp safe to use?
+
Our security agent has analyzed tresor4k/creavores-aeo-mcp and assigned a Trust Score of 95/100 (tier: Verified). See the full breakdown of passed checks and flags on this page.
Who maintains tresor4k/creavores-aeo-mcp?
+
tresor4k/creavores-aeo-mcp is maintained by tresor4k. The last recorded GitHub activity is dated 2026-10-12, with 0 open issues.
Are there alternatives to creavores-aeo-mcp?
+
Yes. On ClaudeWave you can browse similar mcp servers at /categories/mcp, sorted by popularity or recent activity.
Deploy creavores-aeo-mcp to your cloud
Ship this repo to production in minutes. Each platform spins up its own environment with editable env vars.
Maintain this repo? Add a badge to your README
Drop the badge into your GitHub README to show it's tracked on ClaudeWave. Each badge links back to this page and reflects the live Trust Score.
[](https://claudewave.com/repo/tresor4k-creavores-aeo-mcp)<a href="https://claudewave.com/repo/tresor4k-creavores-aeo-mcp"><img src="https://claudewave.com/api/badge/tresor4k-creavores-aeo-mcp" alt="Featured on ClaudeWave: tresor4k/creavores-aeo-mcp" width="320" height="64" /></a>More MCP Servers
Reverse engineer anything with agents, from app behavior down to native binaries.
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
An open-source AI agent that brings the power of Gemini directly into your terminal.
Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface
🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ and follow here for daily tips and tricks: https://x.com/Scrapling_dev