atomno-mcp-seo-audit
MCP (Model Context Protocol) server for technical SEO & GEO audits, powered by
the detail.web engine — real measurements, not LLM
guesses. Run it from Cursor, Claude Desktop or any MCP client. 8 tools:
audit_site (deterministic 0–100 health score + letter grade, 78 checks across
8 categories, plus a GEO sub-score — visibility in ChatGPT / Perplexity /
Google AI Overviews), audit_diff (compare vs the previous snapshot),
robots.txt & sitemap.xml validators, JSON-LD & meta/OpenGraph builders, and
per-check fix explainers. Probes TLS, redirects, TTFB, AI-crawler access (GPTBot)
and llms.txt. Free tier + PRO (deep-crawl, GEO, 40+ deeper checks).
Why pair it with an LLM?
A language model on its own infers a site from training data and, at best, one
rendered page — it can't directly read your TLS certificate's expiry, measure
response time, parse sitemap.xml, or check whether GPTBot is blocked in
robots.txt. This server runs those checks for real: actual HTTP requests,
security headers, redirect chains, structured-data validation — and returns a
deterministic score (same site → same number), reproducible enough to put in
a client report. Think of it as the instrument and the LLM as the analyst that
interprets the readout — the two work best together.
What you get
audit_site(url, depth=1, lang="ru")— one call returns:- health score
0–100(higher is better) and a letter gradeA–F; - issues grouped by category (security, SEO & indexing, performance, GEO, …),
each with status
pass / warn / fail; - a short human-readable summary.
- health score
audit_diff(url, lang="ru")— re-audits a site and compares it to the previous run: health/score delta and which checks got worse or better. The first call stores a baseline. This is something a one-off LLM question can't do — track a site over time. Stateful PRO feature (needs an API key).list_checks(lang="ru")— the full catalogue of engine checks grouped by category, with afree/PRObadge on each — so you (and the agent) can see exactly what the free tier covers and what PRO unlocks.explain_issue(check_id, lang="ru")— a deep-dive on a single check: why it matters and how to fix it. Pass acheck_idfromaudit_siteorlist_checks. Title and category are localized; detailed advice is currently in Russian (theadvice_langfield reports this).validate_robots(content, lang="ru")— paste arobots.txtand get back syntax issues, whether aSitemap:directive is present, whether CSS/JS is blocked from render bots, and which AI crawlers (GPTBot, ClaudeBot, …) are explicitly blocked. No fetch — validates the text you provide.check_sitemap(url, lang="ru")— fetches a sitemap by URL and reports its format (urlset/sitemapindex), URL count and common problems (404, non-XML content type,http://links, missing<lastmod>, the 50k-per-file limit). The fetch is SSRF-guarded on the server.build_jsonld(type, fields, lang="ru")— generates a ready-to-paste schema.org JSON-LD<script>(Organization, LocalBusiness, Article, Product, FAQPage, BreadcrumbList, WebSite) and tells you which required/recommended fields are missing. It never invents data — only what you pass in.build_meta(fields, lang="ru")— generates<head>meta tags (title, description, canonical, Open Graph, Twitter Card) and validates the title (50–60 chars) and description (120–160 chars) lengths.
Free vs PRO
| Free (no key) | PRO (with API key) | |
|---|---|---|
| Checks | core technical basics | 40+ deeper checks (E-E-A-T, Schema.org, Goldmine title) |
| GEO | 4 GEO signals | GEO readiness sub-score + deep GEO checks |
| Crawl | single page | deep-crawl up to 20 pages (depth=2/3) |
The audit engine itself stays on the server — this package is a thin client (HTTP calls + formatting only).
Install
uvx atomno-mcp-seo-audit
Or add to your MCP client config (mcp.json):
{
"mcpServers": {
"seo-audit": {
"command": "uvx",
"args": ["atomno-mcp-seo-audit"]
}
}
}
Configuration
All via environment variables:
| Variable | Default | Purpose |
|---|---|---|
DETAILWEB_API_BASE | https://api.detailweb.ru | Backend base URL |
DETAILWEB_API_KEY | — | PRO key (dwa_...). Without it → free tier |
DETAILWEB_TIMEOUT | 60 | HTTP timeout (seconds) |
DETAILWEB_LANG | ru | Default issue-title language (ru / en) |
The free tier needs no key and no signup — just run the command above. The
PRO tier (40+ deeper checks, GEO sub-score, deep-crawl, audit_diff) is
currently provisioned on request: email kir@detailweb.ru or reach out via
audit.detailweb.ru. Once your account is active you
create keys yourself in Dashboard → Account → API keys (dwa_…, shown once)
and put the key in DETAILWEB_API_KEY.
Example
"Audit https://example.com"
The agent calls audit_site("https://example.com") and gets back the health
score, grade and the list of issues to fix.
License
MIT © atomno-labs. The open-source client talks to a proprietary hosted backend.
🇷🇺 На русском
MCP-сервер технического SEO-аудита на движке detail.web. Запускайте аудит прямо из ИИ-агента (Cursor, Claude Desktop и любой MCP-клиент): health-score, проблемы по 8 категориям и GEO-суб-балл (видимость в ИИ-поиске — ChatGPT, Perplexity, AI Overviews).
Зачем в связке с нейросетью. Языковая модель сама по себе судит о сайте по
обучающим данным и в лучшем случае по одной отрисованной странице — она не
прочитает напрямую срок SSL-сертификата, не измерит время ответа, не распарсит
sitemap.xml и не проверит, заблокирован ли GPTBot в robots.txt. Этот сервер
выполняет такие проверки по-настоящему: HTTP-запросы, заголовки, редиректы,
микроразметка — и даёт детерминированный score (тот же сайт → то же число),
пригодный для отчёта клиенту. Это прибор, а нейросеть — аналитик, который читает
показания. Лучше всего работает связка.
Инструменты: audit_site (аудит + score + GEO), audit_diff (что
изменилось с прошлой проверки — stateful PRO), list_checks (каталог
проверок free/PRO), explain_issue (почему важно + как исправить),
validate_robots, check_sitemap, build_jsonld, build_meta.
Установка:
uvx atomno-mcp-seo-audit
В конфиге MCP-клиента (mcp.json):
{
"mcpServers": {
"seo-audit": {
"command": "uvx",
"args": ["atomno-mcp-seo-audit"],
"env": { "DETAILWEB_LANG": "ru" }
}
}
}
Бесплатный тариф (базовые проверки, одна страница) работает сразу, без ключа
и регистрации. PRO-режим (40+ глубоких проверок, GEO-суб-балл, deep-crawl до
20 страниц, audit_diff) пока выдаём по запросу: напишите на
kir@detailweb.ru или через audit.detailweb.ru.
После активации аккаунта ключ (dwa_…) создаётся в кабинете → Аккаунт →
API-ключи (показывается один раз) и подставляется в DETAILWEB_API_KEY в
env. Полное описание инструментов и настроек — в английской версии выше.