lab — working exhibits

Run the instrument.

Exhibits, not screenshots. The flagship grades any site's readiness for the machines doing the reading — ChatGPT, AI Overviews, Perplexity, agents — live, in full, ungated. Point it at your site. Point it at this one.

10 exhibits installed · scoring model published · agents welcome via mcp

exhibit 01 — agent-ready grader

live engine · ungated · 8 runs/hr per visitor

exh-10 — the deposition · the site cross-examines itself

13 surfaces · 11 facts · re-examined every ten minutes

01the operator's namestated on 12 of 13 · Jamie McKayeconsistent

/ (html)Jamie McKaye
/about/ (html)Jamie McKaye
/work/ (html)Jamie McKaye
/services/ (html)Jamie McKaye
/writing/ (html)Jamie McKaye
/contact/ (html)Jamie McKaye
json-ld graph (in every route)Jamie McKaye
/md/ (markdown)Jamie McKaye
/md/about/Jamie McKaye
/md/work/—
/llms.txtJamie McKaye
/feed.xml (rss)Jamie McKaye
mcp · tools/listJamie McKaye

02in practice sincestated on 6 of 13 · 2007consistent

/ (html)2007
/about/ (html)2007
/work/ (html)—
/services/ (html)—
/writing/ (html)—
/contact/ (html)—
json-ld graph (in every route)2007
/md/ (markdown)2007
/md/about/2007
/md/work/—
/llms.txt2007
/feed.xml (rss)—
mcp · tools/list—

03years in the fieldstated on 5 of 13 · 19consistent

/ (html)19
/about/ (html)19
/work/ (html)—
/services/ (html)19
/writing/ (html)—
/contact/ (html)—
json-ld graph (in every route)—
/md/ (markdown)19
/md/about/19
/md/work/—
/llms.txt—
/feed.xml (rss)—
mcp · tools/list—

04the direct linestated on 4 of 13 · jamie@jamiemckaye.comconsistent

/ (html)jamie@jamiemckaye.com
/about/ (html)—
/work/ (html)—
/services/ (html)—
/writing/ (html)—
/contact/ (html)jamie@jamiemckaye.com
json-ld graph (in every route)jamie@jamiemckaye.com
/md/ (markdown)—
/md/about/—
/md/work/—
/llms.txtjamie@jamiemckaye.com
/feed.xml (rss)—
mcp · tools/list—

05where the practice sitsstated on 5 of 13 · Hershamconsistent

/ (html)Hersham
/about/ (html)—
/work/ (html)—
/services/ (html)—
/writing/ (html)—
/contact/ (html)Hersham
json-ld graph (in every route)Hersham
/md/ (markdown)Hersham
/md/about/Hersham
/md/work/—
/llms.txt—
/feed.xml (rss)—
mcp · tools/list—

06clients named with permissionstated on 5 of 13 · Nike · Pitney Bowes · Urban Outfittersconsistent

/ (html)Nike · Pitney Bowes · Urban Outfitters
/about/ (html)—
/work/ (html)Nike · Pitney Bowes · Urban Outfitters
/services/ (html)Pitney Bowes · Urban Outfitters
/writing/ (html)—
/contact/ (html)—
json-ld graph (in every route)—
/md/ (markdown)—
/md/about/—
/md/work/Nike · Pitney Bowes · Urban Outfitters
/llms.txtNike · Pitney Bowes · Urban Outfitters
/feed.xml (rss)—
mcp · tools/list—

07the one verified numberstated on 4 of 13 · 2.2 s → 47 msconsistent

/ (html)2.2 s → 47 ms
/about/ (html)—
/work/ (html)2.2 s → 47 ms
/services/ (html)—
/writing/ (html)—
/contact/ (html)—
json-ld graph (in every route)—
/md/ (markdown)—
/md/about/—
/md/work/2.2 s → 47 ms
/llms.txt2.2 s → 47 ms
/feed.xml (rss)—
mcp · tools/list—

08recordings re-instrumentedstated on 4 of 13 · 31consistent

/ (html)31
/about/ (html)31
/work/ (html)—
/services/ (html)—
/writing/ (html)31
/contact/ (html)—
json-ld graph (in every route)—
/md/ (markdown)—
/md/about/—
/md/work/—
/llms.txt31
/feed.xml (rss)—
mcp · tools/list—

09the corpus retired at the resetstated on 6 of 13 · 222consistent

/ (html)222
/about/ (html)222
/work/ (html)222
/services/ (html)222
/writing/ (html)222
/contact/ (html)222
json-ld graph (in every route)—
/md/ (markdown)—
/md/about/—
/md/work/—
/llms.txt—
/feed.xml (rss)—
mcp · tools/list—

10mcp tools on offerstated on 5 of 13 · ask_concierge · contact_channel · grade_site · request_engagement · site_mapconsistent

/ (html)ask_concierge · contact_channel · grade_site · request_engagement · site_map
/about/ (html)—
/work/ (html)—
/services/ (html)—
/writing/ (html)—
/contact/ (html)request_engagement
json-ld graph (in every route)—
/md/ (markdown)ask_concierge · contact_channel · grade_site · request_engagement · site_map
/md/about/—
/md/work/—
/llms.txtask_concierge · contact_channel · grade_site · request_engagement · site_map
/feed.xml (rss)—
mcp · tools/listask_concierge · contact_channel · grade_site · request_engagement · site_map

11pieces on the recordstated on 2 of 13 · 32consistent

/ (html)—
/about/ (html)—
/work/ (html)—
/services/ (html)—
/writing/ (html)32
/contact/ (html)—
json-ld graph (in every route)—
/md/ (markdown)—
/md/about/—
/md/work/—
/llms.txt—
/feed.xml (rss)32
mcp · tools/list—

the finding

100%

agreement across 11 facts stated on two or more surfaces · 0 contradictions

Agents don't read a page; they depose a site — every surface, against every other. This exhibit runs that deposition on jamiemckaye.com itself, over HTTP, exactly as an agent would: the pages, the structured data, the markdown mirrors, the manifest, the feed, the MCP tool list. Plain rules, no model. A contradiction here is a bug, and the bug is public.

sat 2026-09-15 19:08 utc · 13 of 13 surfaces answered · slowest 384 ms

the argument — your website will be deposed →

exh-05 — the readers · the machine guestbook

page reads by known crawlers & agents · rolling 7 days

Meta-External21,591

Googlebot1,408

Bingbot624

ChatGPT-User189

Amazonbot154

ClaudeBot111

CCBot100

Bytespider72

OAI-SearchBot70

Facebook-Preview68

Claude-User66

YandexBot64

why this exists

The thesis of this site is that search multiplied — Google, answer engines, agents. This wall is that claim, measured on this site's own traffic: who reads, how often, by name.

exh-06 — grader league · every public scan, anonymised

15 scans · rolling 30 days

11A

the loop · plus this site's own pulse

Scans feed the distribution; the distribution calibrates the grade. Nothing identifying is stored — a salted hash joins the league, the domain never does.

visitors' p75 this week — lcp 1.24s · inp 72ms · cls 0.001 · 1000 samples

the bench — what's installed

EXH-01

agent-ready grader

grades any site's readiness for machine readers — engine ported from production, scoring model published

live

EXH-02

mcp endpoint

this site as a tool server — agents can call the grader directly

live — POST /api/mcp

EXH-03

llms.txt

the machine-readable map of this site, served at /llms.txt

live

EXH-04

concierge

ask the site about the work — live in the /services/ai-systems terminal, and via mcp as ask_concierge

live

EXH-05

the readers

the machine guestbook — every crawler and agent that reads this site, counted and credited below

live — counting

EXH-06

grader league

every public scan joins an anonymous distribution — domains hashed, grades kept

live

EXH-07

markdown mirrors

every route and every piece served as clean markdown at /md/ — or ask any URL with Accept: text/markdown and the mirror answers in place

live — GET /md/

EXH-08

vitals loop

real-user web vitals, sampled and aggregated — the footer shows yours, this bench shows everyone's

live — measuring

EXH-10

the deposition

the site cross-examines itself every ten minutes — every surface fetched over http, facts extracted by plain rules, cross-surface agreement scored. a contradiction here is a public bug

live — every 10 min

EXH-09

a11y audit

axe-core + lighthouse, wcag 2.2 aa, all signed pages — three violations found across two audits (hero split aria; ⌘K label-in-name; lens scroll region focus), each fixed same day; zero open

clean — re-audited 01 sep 2026

mcp: streamable http at /api/mcp — tools: grade_site · site_map · ask_concierge · request_engagement · contact_channel. llms.txt at /llms.txt · markdown mirrors at /md/.