4
Content chunking, data as citation bait, off-site brand mentions, and query fan-outs.
Boston-based SEO consultant. I fix search problems and build the tools that prevent them.
Jake Labate is a Boston-based SEO consultant who treats search as infrastructure rather than as a campaign.1
He sells engagements scoped as named deliverables, and publishes the nine-discipline methodology those engagements run on.2 He also builds and operates his own software, including a free universal WebMCP connector that makes a page's forms usable by an agent.34
Recent releases include 20 Claude SEO skills on npm.5
Nine disciplines in three pillars that turn advanced SEO into a reasoning system: how decisions get made, how the site expresses meaning, and how the work gets executed.
Everything I build splits the work that way. In my 20 Claude SEO skills, code runs every crawl, every parse, and all 200 checks. Claude only makes the calls a rule cannot, like writing the fix or choosing between options. You set the scope and decide what ships.
Swipe sideways to follow the steps in order.
Set Scope
You: The site, the page cap, and any thresholds or business context the checks need.
Crawl
Code: Fetches every page once into a shared cache.
fetch_pages.py→page_cache.json
Extract
Code: Parses only the fields one audit needs. No network calls.
extract_*.py→inventory.json
Run Checks
Code: Applies the checks and thresholds and assigns every finding a severity.
audit_*.py→audit_report.json
Judge And Write
Claude: Makes the calls a rule cannot: the replacement copy, the right consolidation, the order to fix things in. Every value traces back to the audit data; nothing is invented.
Export
Code: Turns the report into a Word doc and the findings into a CSV.
md_to_docx.pyfindings_to_csv.py
Decide And Ship
You: Confirm what only you can know, such as intent and priorities, and own what ships.
I can’t stop building stuff. Send help.
SchemaCDNLive
JSON-LD generated per URL and injected at page load.
AuditBuilderLive
Scored findings, exported as a PDF, a Sheet and an API.
PageSpeed SamplerLive
A sampled Lighthouse report for a whole site, per collection.
WhyThoLive
Notes pinned to page elements by selector.
Open GraphLive
A 1200 by 630 share image and its meta tags, from one URL.
Context CardsLive
Hover definitions for the terms in a page's DefinedTermSet.
Page ShareLive
A floating share button that builds each platform's link.
Client PortalLive
One coded address per client, holding every deliverable.
GroundTruthLive
One question card at a time, filed as evidence for beliefs.
Deploy LayerLive
SEO settings published to live pages, every change undoable.
1Soon
An append-only ledger of claims, each tagged with how it was made.
AISurfaceSoon
One row per AI answer: named, position, tone, citations.
GapFinderSoon
Competitor claims your site lacks, with the source text.
AnalyticsSoon
Search Console and GA4, traced from total to query and page.
BrandWieldSoon
A sourced business record, and posts drafted only from it.
Schema markup is slow, expensive and fragile to implement by hand. SchemaCDN replaces page-by-page JSON-LD with a single script tag: the page is read, the data is mapped to schema.org by rules, stored per URL, and delivered at runtime.
<script
defer
src="https://deploy.schemacdn.com">
</script>
npm i claude-seo-skillsClaude works better with the Skills feature, so I built 20 of them, each a different SEO use case, and made them available for free download. Each Claude SEO Skill defines inputs, guardrails, produced results.
AI agents don’t only read the web, but interact with it. The WebMCP protocol allows for just this. I built a free, instant, universal WebMCP connector.
WebMCP asks you to hand-annotate every form on every template, and that decays the first time someone ships a redesign. So I built the framework that does it at runtime instead: one script walks the DOM, finds every form, and writes the attributes an agent needs to invoke it as a tool.
<script
defer
src="https://webmcp.schemacdn.com">
</script>
Ten of the loudest GEO claims, tried against first-party documentation and published research. Most of what gets asserted survives only as plausible.
4
Content chunking, data as citation bait, off-site brand mentions, and query fan-outs.
5
Schema markup, llms.txt, rendering strategy, markdown mirrors, and AI crawler control in robots.txt.
1
Selective content gating, which withholds the exact evidence AI systems need in order to cite you.
A+ technical talent.
The depth of his knowledge is breathtaking.
Jake Labate is a master at his craft.
He went above and beyond for the work he did for me.
He is thorough and pays attention to every detail.
[He] is extremely knowledgeable and tech savvy.
Jake stands out as an expert who seamlessly merges creative design and strategic SEO practices.
He's an excellent mentor and role model.
Jake Labate's marketing services are first rate.
He was very knowledgeable on all facets of online marketing.
Best person I've gone to! Hands down! 10/10 would recommend.
When it comes to SEO, no one bests Jake.
Jake has a strong understanding of the latest search algorithm updates and trends.
You're lowkey teaching me how to communicate, not just telling me everything is fine.
You remind me of someone who will continue to move on and on to better things. Everytime I see you, it makes my day, I don't know why.
You're the first person to tell me ‘why wait’.
Three engagements, each measured the same way, across 5M+ enterprise pages managed and analyzed.
+62.5%
Indexing recovered
800k to 1.3M indexed pages on a large auto marketplace, driven by technical-first execution that held long after launch.
+200%
Traffic multiplied
800 to 2,400 monthly visits for an investment-focused client, with the measurement discipline to prove which lever did it.
+14%
SERP CTR from schema
A hotel-industry schema framework that worked well enough to become SchemaCDN.
Indexing that won't recover, schema that drifts, or a search program that stalled. Say which one and I'll tell you whether I'm the right person for it.