MCP Hub
Back to servers

mcp

AEO scoring, llms.txt audit, and agent-readiness checks for AI agents.

Registry
Updated
May 10, 2026

Quick Install

npx -y @neverranked/mcp

Never Ranked

AI-native SEO & AEO agency. Search changed. Your visibility didn't. We fix that.

This repository contains the Never Ranked marketing site, the audit methodology used to deliver client work, and the first proof-of-practice case study (an audit of our own in-house SaaS, Montaic).


What's in this repo

neverranked/
├── index.html                  # the marketing site (single-file, no framework)
├── _headers                    # Cloudflare Pages security + caching headers
├── robots.txt                  # robots directives
├── sitemap.xml                 # sitemap for crawlers
│
├── audit-template/             # reusable blank template for client audits
│   ├── README.md               # how to run a new audit
│   ├── intake-questionnaire.md # 12-question client intake
│   ├── 00-executive-summary.md
│   ├── 01-intake.md
│   ├── 02-technical-audit.md
│   ├── 03-schema-review.md
│   ├── 04-keyword-gap.md
│   ├── 05-ai-citations.md
│   ├── 06-competitor-teardown.md
│   └── 07-roadmap.md
│
├── audits/                     # completed client audits
│   └── montaic/                # the first case study
│       ├── 00-executive-summary.md
│       ├── 02-technical-audit.md
│       ├── 03-schema-review.md
│       ├── 04-keyword-gap.md
│       ├── 05-ai-citations.md
│       ├── 06-competitor-teardown.md
│       ├── 07-roadmap.md
│       ├── implementation/     # Month 1 pasteable implementation kit
│       │   ├── README.md
│       │   ├── A1-root-schema.md
│       │   ├── A2-breadcrumbs.md
│       │   ├── A3-howto-schema.md
│       │   ├── A4-blogposting-schema.md
│       │   ├── A5-software-application.md
│       │   ├── A6-free-grader-full-stack.md
│       │   ├── A7-canonicals-robots.md
│       │   ├── A8-og-images.md
│       │   ├── A9-title-meta-rewrites.md
│       │   ├── A10-entity-registration.md
│       │   └── A11-fair-housing-pillar-article.md
│       └── raw/                # fetched HTML + evidence trail
│
└── scripts/
    └── run-audit.py            # automated audit runner (no dependencies)

The site

index.html is the entire Never Ranked marketing site. Single file. No framework. No build step. Editorial dark-luxury design inspired by MouthWash Studio.

Fonts: Playfair Display (headlines), DM Mono (body), Barlow Condensed (labels). Palette: #080808 background, #c9a84c gold accents, #f0ece3 off-white text. Features: Film grain overlay, scroll reveal animations, mobile-first responsive, async-first positioning.

Deploy: drop this folder into Cloudflare Pages, or point a GitHub connection at it. No build command. Output directory: /.


The methodology

Never Ranked delivers audits in six parts:

  1. Technical Audit — meta tags, canonicals, OG tags, schema inventory, crawlability
  2. Schema Review — JSON-LD coverage, entity signals, ready-to-paste fixes
  3. Keyword Gap Analysis — SERP reality check, intent clusters, positioning opportunities
  4. AI Citation Audit — citation share across target queries in ChatGPT, Perplexity, Gemini, Google AI Overviews
  5. Competitor Teardown — side-by-side technical comparison with direct competitors
  6. 90-Day Roadmap — prioritized, dated, effort-estimated action plan

The audit runs in ~90 minutes of automated work plus ~3 hours of human synthesis. Delivered within 48 hours of booking.

Running a new audit

# 1. Clone this repo, cd into it
git clone https://github.com/yourorg/neverranked.git
cd neverranked

# 2. Run the audit runner against the target domain
python3 scripts/run-audit.py https://client-domain.com --out audits/client-name/raw/

# 3. Copy the template
cp -r audit-template audits/client-name

# 4. Fill in each deliverable using the raw data + live SERP testing
# (see audit-template/README.md for the full workflow)

The Montaic audit

The audits/montaic/ directory contains the first Never Ranked audit, run against our own in-house SaaS. It's the proof of practice — we don't pitch what we haven't tried.

Headline finding: Montaic has 0% AI citation share across 8 primary category queries, vs 100% for ListingAI (the category leader).

Root cause: Missing entity signals. No Organization schema anywhere on the site, despite 222 indexed pages. Google fuzzy-matches "Montaic" as "Monti" / "Monte" / "Montana."

Fix path: The implementation/ subdirectory contains pasteable code and copy for all 11 Month 1 tasks. Total implementation time: ~10 focused hours spread across 3-5 days.

Expected outcome: 15%+ AI citation share within 90 days, paired with Wikidata / Crunchbase / G2 / LinkedIn entity registration.


The operating model

Never Ranked is deliberately small. One founder, async-first, no standing meetings.

  • Audit: $500 one-time, delivered within 48 hours. The starting diagnostic.
  • Signal retainer: $2,000/mo, three-month minimum. Web-grounded citation tracking across ChatGPT, Perplexity, Gemini, Claude. Reddit thread monitoring. Schema fixes auto-pushed to your live site (graded for completeness first). Authority audits. Industry-percentile benchmarks. Quarterly drift detection. Forward-ready Monday digest.
  • Amplify retainer: $4,500/mo, three-month minimum. Capped at 6 active clients globally. Everything in Signal plus brand-voice fingerprint, citation-shaped content drafted in your voice, auto-publish to WordPress / Webflow / Shopify, and Reddit reply briefs (per-thread strategic briefs your team uses to post real human replies — we never draft the comment itself).

The dashboard is the meeting. Loom recaps replace calls. Email SLAs replace Slack channels.


Contact

Email: hello@neverranked.com Web: https://neverranked.com Proof of practice: Montaic — our in-house AI-native SaaS, audited by the same team that runs Never Ranked.


License

This repository is source-available, not open-source. See LICENSE for the full text. In plain language:

  • You can: read every line, audit it, run it locally, fork it, learn from it, cite it, file issues, send PRs, copy ideas into your own work.
  • You cannot: use this code to operate a commercial AEO service (citation tracking, schema grading, drift detection, authority audits, etc.) that competes with NeverRanked.
  • In four years (April 2030): the code automatically converts to MIT and the restriction lifts.

The license is modeled on the Business Source License pattern used by MariaDB, Sentry, and CockroachDB. The intent: keep the code open enough that customers can audit our work and security claims, while preventing a clone-and-launch competitor in the early years.

If you want to operate a commercial service using this code before the change date, email licensing@neverranked.com.

The "NeverRanked" name and wordmark are trademarks. Forks must be renamed.

Reviews

No reviews yet

Sign in to write a review