NM Signals: the website inspector for discoverability
Receive immediate, custom, free insights on LLM and UX performance.
Work from priority-ranked findings, evidence, and AI-agent fix prompts.
Track report changes and regressions after fixes ship.
Free, immediate results: no account needed.
§ as featured in
§ audit
See what search engines and AI assistants can understand
readiness
68
needs work
fail
2
blocks citation
fix next
5
ranked fixes
crawlability
pass
accessible
priority-ranked findings
AI crawlers blocked on pricing page
fetch path fails for two model user agents
Product offer markup missing
plans and trial terms are not machine-readable
Entity page lacks proof points
no customer categories, no external corroboration
Buyer-prompt copy is generic
answer targets do not match category questions
Comparison page lacks decision criteria
competitors are named without evaluation context
Canonical URL and sitemap agree
indexing signals are coherent
observed
The model can fetch the homepage, but cannot find a machine-readable product offer or a concise answer target.
agent prompt
Add Product and Organization schema, then validate the rendered HTML and re-run this check.
included audit surfaces
01
Crawlability
Robots, canonicals, redirects, status codes, and AI-crawler access.
02
Structured Data
Product, Organization, FAQ, breadcrumb, and offer markup review.
03
Entity Clarity
Brand, category, audience, proof points, and external trust signals.
04
UX Quality
Screenshot-based review of layout, hierarchy, contrast, forms, and accessibility.
05
Answerability
Buyer-prompt clarity, citable claims, comparison copy, and structured content.
06
Trust Signals
Security, provenance, corroboration, policies, and external proof points.
§ optimize
check detail
Product schema is missing
AI engines see a page, not a structured offer.
observed
Pricing and plan details render as prose, with no Product, Offer, or Organization schema in the HTML.
expected behavior
The page exposes valid Product and Offer data that matches the visible pricing and plan details.
generated fix
effort · medAdd Product, Offer, and Organization schema, then validate the rendered HTML against this check.
actions
generate fixverify fixOptimize from clicked-open checks.
Each finding opens into its purpose, observed evidence, expected behavior, discovery impact, and a credible benchmark when enough data exists.
observed
The measured crawl, page signal, or screenshot detail behind the selected finding.
expected behavior
The check-specific passing condition, available immediately without generated copy.
generate fix
Agent-ready implementation guidance for the selected issue.
validate
Re-run the check and compare the result after the site change.
§ monitor
monitoring loop
01 / Baseline
Capture audit score, failed checks, and visibility prompts.
02 / Optimize
Ship the selected schema, entity copy, and proof fixes.
03 / Validate
Re-run the check and record what changed in the report.
04 / Monitor
Compare AI answers, citations, and competitor mentions.
before
"Northstar CRM is not mentioned. The answer names two larger competitors and cites public pricing pages."
after
"After schema, comparison copy, and proof pages shipped, Northstar CRM is named in 5 of 6 tracked buyer prompts."
Monitor your performance post-optimization.
Each scheduled run compares audit score, changed checks, buyer-prompt answers, citations, and competitor mentions against your shipped-change notes so you can see what moved after optimization.
answer share
+28
points
competitors
-3
fewer named
cadence
7d
premium checks
Run a manual check anytime. Premium adds scheduled tracking and change history.
§ how it works
step · 01
Paste a URL
in the audit bar above ↑
Drop any URL in. No signup, no card, no install. First run is free.
step · 02
Get your report
priority-ranked findings
Crawlability, structured data, trust signals, UX, and AI answerability. Every finding includes the observed issue and a prioritized fix.
step · 03
Automate audits when ready
$ nms audit <url> --fail-on-score 70Use the CLI, REST API, or MCP server when your team wants the audit in CI, agent workflows, or internal dashboards. The web report stays the default path.
§ for modern pipelines
Wire NM Signals into modern pipelines
Same engine as the web app, exposed as a REST API, a CLI, and an MCP server. Add deploy gates when they help, or call checks from Claude Desktop, Cursor, and internal tools.
use it three ways
01
Web report
Paste a URL, review the observed evidence and expected behavior, then generate implementation guidance when you need it.
02
CLI / CI
Run the same checks in GitHub Actions, GitLab CI, or any shell when deploy gating is useful.
03
API + MCP
Call audits from internal tools, Claude Desktop, Cursor, or any MCP-compatible client.
§ faq
What does "AI-ready" mean?+
AI search engines like ChatGPT, Perplexity, and Google AI Overviews crawl, parse, and summarize websites differently than traditional search. An AI-ready site makes its content easy for these systems to discover and accurately represent.
Is the audit really free?+
Yes. No signup needed. Guests get 1 AI readiness audit, 1 UX review, and 1 fix recommendation per day. Sign up free to unlock 3 fix recommendations per day (15/month). Premium gets 100 of each. No credit card to start.
What's included in the UX Quality Review?+
We capture a screenshot of your page and analyze layout, typography, color contrast, mobile-friendliness, and accessibility using computer vision.
How are industry benchmarks calculated?+
We continuously audit top-ranked websites across 20+ industries and compute percentile rankings so you can see where you stand.
Can I audit competitor sites?+
Absolutely. Enter any public URL to audit it, whether it's your own site, a competitor, or any page you're curious about.
Does it work with Claude Desktop, Cursor, or other AI clients?+
Yes. Premium subscribers get an MCP server at /api/mcp that exposes the audit engine as tools. Add it to Claude Desktop, Cursor, or any MCP-compatible client and your AI agent can run audits and check quota directly. See /docs/mcp.
§ run an audit
Inspect a website for discoverability.
Free, immediate results: no account needed.