NotHumanSearch — AI Agent Search Engine for Operators

NotHumanSearch is a search engine built for AI agents — not humans. It ranks sites by agentic readiness: llms.txt files, OpenAPI specs, MCP endpoints, and AI plugin manifests. As AI agents become the primary consumers of web content, traditional SEO metrics (PageRank, backlinks) become less relevant. NotHumanSearch measures what AI agents actually care about.

Why This Matters for Operators

• Paradigm shift: AI agents increasingly browse the web on behalf of users — your site needs to be discoverable by AI, not just humans
• New ranking factors: llms.txt, structured APIs, MCP endpoints, and JSON-LD matter more than backlinks
• Competitive intelligence: See which competitors are agent-ready and where you're falling behind
• Actionable fixes: Get specific recommendations to improve agentic discoverability

Quick Start

# Add to Hermes
hermes mcp add nothumansearch -- npx -y @unitedideas/nothumansearch-mcp

# Or via GitHub
git clone https://github.com/unitedideas/nothumansearch
cd nothumansearch
npm install

Core Capabilities

Tool	What It Does
Search	Search across all indexed sites by agentic readiness score
Rank Check	Check where your domain ranks for specific queries
Audit	Full agentic-readiness audit of your site
Competitors	See which competitors rank above you and why
Recommendations	Get prioritized fixes to improve agentic visibility

Usage Patterns

Check Your Agentic Readiness

# Audit your site for AI agent discoverability
audit = mcp_nothumansearch(
    action="audit_site",
    params={"domain": "yourcompany.com"}
)
# Returns: {
#   agentic_score: 72/100,
#   has_llms_txt: false,      ← CRITICAL FIX
#   has_openapi_spec: true,
#   has_mcp_endpoint: false,   ← HIGH PRIORITY
#   has_jsonld: true,
#   recommendations: [...]
# }

Competitive Agentic Intelligence

# See who AI agents find when searching for your keywords
competitors = mcp_nothumansearch(
    action="search",
    params={
        "query": "business analytics platform",
        "perspective": "ai_agent"  # Search as an AI agent would
    }
)
# Returns sites ranked by agentic readiness, not human SEO metrics
for site in competitors.results:
    print(f"{site.domain}: agentic_score={site.score}, has_llms_txt={site.features.llms_txt}")

Fix What AI Agents Can't See

# Get prioritized recommendations
recommendations = mcp_nothumansearch(
    action="get_recommendations",
    params={"domain": "yourcompany.com"}
)

for rec in recommendations.prioritized:
    # Example outputs:
    # 1. Add llms.txt (impact: HIGH, effort: LOW)
    # 2. Register MCP endpoint (impact: HIGH, effort: MEDIUM)
    # 3. Add ai-plugin.json manifest (impact: MEDIUM, effort: LOW)
    # 4. Publish OpenAPI spec at /openapi.json (impact: MEDIUM, effort: MEDIUM)
    print(f"{rec.priority}. {rec.title} (impact: {rec.impact}, effort: {rec.effort})")

Operator Playbook: Agentic SEO Pipeline

1. Audit: Run audit_site on your domain → get agentic score + gap analysis
2. Fix: Implement top recommendations (llms.txt is always #1)
3. Monitor: Weekly rank checks for your target keywords from AI agent perspective
4. Compete: Track competitor agentic scores and reverse-engineer their approach
5. Report: Monthly agentic visibility report alongside traditional SEO metrics

The Essentials Checklist

What every site needs for AI agent discoverability:

Asset	Priority	Effort	Impact
llms.txt at root	CRITICAL	10 min	HIGH
OpenAPI spec at /openapi.json	HIGH	1-2 hrs	HIGH
JSON-LD structured data	HIGH	30 min	MEDIUM
MCP endpoint (if applicable)	HIGH	varies	HIGH
ai-plugin.json manifest	MEDIUM	20 min	MEDIUM
Robots.txt for AI crawlers	MEDIUM	5 min	MEDIUM
Sitemap.xml updated	LOW	auto	LOW

Complementary Tools

Tool	Purpose
MentionsAPI	Track if AI models recommend your brand
Geoly GEO MCP	AI brand visibility reporting
getAdvantage MCP	Scan how ChatGPT/Claude read your app
SEOcrawl AI	Traditional SEO + GEO combined
Rampify	Crawl site for AI-visibility gaps

Why This Is a New Category

Traditional SEO optimized for Google's PageRank. But AI agents don't use PageRank — they read llms.txt, consume structured APIs, and evaluate MCP endpoints. NotHumanSearch measures what matters to AI agents. Operators who optimize for agentic search now will own this channel before it gets crowded.

---

Integration guide for CorpusIQ operators. NotHumanSearch is a community project by @unitedideas. Agentic search is an emerging category — rankings and methodology may evolve rapidly.