Goblin-run AI visibility

FULLY AUTOMATED SEO AND AI VISIBILITY.

Prompt Goblin finds where ChatGPT, Perplexity, Gemini, Claude, and Google AI answers miss you — then a real specialist fixes the pages, content, and signals that help those engines understand your business.

No auto-posting. No black box. Human-reviewed work, shipped and measured.

Visibility Spellbook3 schools · 1 goblin

AEO

answer-engine visibility

Get cited inside ChatGPT, Claude, Gemini, Perplexity & AI Overviews.

SEO

technical foundation

Crawl, index, schema, Core Web Vitals. The base AEO stands on.

A11Y

WCAG 2.1 AA + Section 508

Accessible to people and parseable by machines. Required for gov.

> one stack: get found, stay legible, stay compliantv2

▸$ goblin engine --explainself-healing · eval-gated · engineer-approved

how it works · scan, fix, ship, measure

How we actually move the needle

// detect your stack · write actual code fixes + assets · you approve · we ship

Most SEO shops send you a PDF of recommendations. We write the actual code.

Under the hood: agentic workflows detect your tech stack, then find the gaps across answer-engine visibility, technical SEO, and accessibility — where ChatGPT, Claude, Perplexity, and Gemini cite competitors instead of you. Bounded self-healing loops and an eval gate prove a fix actually works before any engineer sees it. A software engineer reviews every change, and you get actual code fixes — merge-ready diffs scoped to your stack, not a list of suggestions — plus the branded assets to post. You approve before anything ships to your repo or CMS. Then it re-runs on a schedule and measures the delta.

You see the gap. Then you watch it close.

"RAG pipeline" and "CI/CD eval gate" are the accurate technical names for what runs. We surface them once for credibility, then translate them for everyone else.

goblin-engine · self-healing · eval-gatedsample run

01 · diagnose

Find the gap

RAG pipeline samples engines + audits SEO / a11y, surfaces a scoped fix.

02 · eval gate

Prove it passes

CI/CD eval gate runs the suite. Red on regression → the fix bounces back to self-heal.

○ idle

03 · engineer

An engineer approves

Goblin Mode: every change halts here. A software engineer approves → reviewed PR. Autopilot Mode (upgrade): low-risk content posts on schedule; you choose the mode.

engineer-gated

fix packet · a11y

▸scan → gap found: contrast on .cta (sample)

active pathfailed gate → self-healengineer review gate

▸ One pipeline: diagnose → eval gate → engineer approve. A failed gate bounces the fix back to the self-healing loop; nothing ships until a software engineer approves. SEO + accessibility proven on the gate (2026-06-02); schema scaffolded.

▸$ goblin graph --runlanggraph workflow · engineer-gated · sample run

01
Listen to prompt surfaces
Agents sample real buyer queries across ChatGPT, Claude, Gemini, and Perplexity — plus your Bing rank, the grounding Copilot and Google AI Overviews lean on. Not just keyword tools.
02
Retrieve & diff citation graph
Map who LLMs actually cite for your category, then diff against your domain to expose exactly which sources you're losing to.
03
Audit: schema · SEO · accessibility
One pass flags missing entities & structured data, technical-SEO leaks, thin content, AND WCAG 2.1 AA / Section 508 gaps. Across real rendered component states, not a single snapshot.
04
Goblin recommendation engine
Each gap becomes a ranked, scoped task with impact, effort, and a paste-ready fix prompt a coding agent can act on.
05
Engineer-reviewed PRs
Goblin Mode (default): a software engineer approves every change before it hits your CMS, schema, or repo. Autopilot Mode (upgrade on Mention Engine and Goblin Forge): low-risk content posts on schedule. You choose the mode.
06
Loop on cadence, automatically
The graph re-runs on a schedule and reports the measured before/after delta. Visibility, citations, SEO, and a11y coverage become a tracked KPI, not a vibe.

//telemetrythis site · dogfooded

Answer engines measured

JSON-LD blocks · this site

Pipeline nodes · engineer-gated

Changes auto-deployed

AI search visibility▪cited by ChatGPT▪show up in Perplexity▪answer-engine ready▪get found by robots▪stay usable by humans▪llms.txt▪JSON-LD▪RAG-ready▪structured data▪entity SEO▪Visible AF▪AI search visibility▪cited by ChatGPT▪show up in Perplexity▪answer-engine ready▪get found by robots▪stay usable by humans▪llms.txt▪JSON-LD▪RAG-ready▪structured data▪entity SEO▪Visible AF▪AI search visibility▪cited by ChatGPT▪show up in Perplexity▪answer-engine ready▪get found by robots▪stay usable by humans▪llms.txt▪JSON-LD▪RAG-ready▪structured data▪entity SEO▪Visible AF▪

▸$ man goblinsix services · one goblin

(i)Technical SEO

The plumbing: crawl paths, indexation, canonicals, robots, sitemaps. Fixed so nothing leaks crawl budget.

Crawl audits
Indexation
Canonicals
Robots & sitemaps

(ii)Schema & structured data

JSON-LD so crawlers and models parse who you are without guessing. Table-stakes hygiene: necessary, not magic. The citation levers are brand mentions and Bing rank; this clears the way for them.

JSON-LD
Entity markup
FAQ / HowTo
Rich results

(iii)AI / answer-engine SEO

Get surfaced inside ChatGPT, Perplexity, and AI Overviews by earning the brand mentions and Bing-rank signals that drive citations, plus the llms.txt/AEO hygiene that lets them through.

llms.txt
AEO strategy
Citation tuning
RAG-readiness

(iv)Core Web Vitals

Make it fast for humans and bots alike. Green vitals, real-device tested.

LCP / CLS / INP
Asset budgets
Edge & caching
Lab + field

(v)Content for robots + humans

Pages that read well to a person and parse cleanly for a model. Both audiences, one draft.

Info architecture
Heading logic
Internal links
Editorial passes

(vi)Accessibility (WCAG + 508)

Usable by people on assistive tech and legible to crawlers. The same fixes serve both. Automated axe-core audit across real component states (collapsed, open, error) plus a software-engineer pass, since tooling alone catches ~57%. Required for government (Section 508 / ADA Title II); never sold as compliance-by-tool.

WCAG 2.1 AA + Section 508
Stateful axe-core audit
Engineer-reviewed remediation
Reviewed fix PRs

▸$ goblin --pricingmanaged work · human-approved · monthly

new

$99 Scan

See what AI engines say about you — and what they don't.

// founders testing the waters

$99/ mo

See what ChatGPT, Perplexity, Gemini, and Claude say about you, what they miss, and which competitors show up instead — a human-reviewed weekly report, not a raw export.

+ Human-reviewed weekly visibility report
+ No dashboard login or prompts to configure
+ ChatGPT · Claude · Gemini · Perplexity + your Bing rank
+ Ranked what's-missing list + week-over-week deltas
+ Competitor analysis — who's getting cited instead of you

Book a scan →

$499 Fix

The goblins generate the fixes, assets, and content.

// founders who want the work done for them

$499/ mo

For teams ready to act: everything in Scan, plus reviewed fixes, content, page improvements, and a focused office hour to move the first priorities forward.

+ Everything in Scan
+ Managed AI-audit — 4 answer engines + Bing rank
+ Engineer-reviewed fix queue · scored by impact × effort
+ Generated fixes — schema, copy, content rewrites
+ AI-ready content calendar with prioritized actions
+ Stack-specific snippets + AI-ready prompt artifact
+ 60-min goblin office hour

Start fixing →

best valueMost teams start here

$999 Mention Engine

Grows your AI visibility — a goblin reviews every fix before it ships.

// teams that want ongoing momentum, not a report

$999/ mo

For teams that want the monthly loop handled: we scan, prioritize, create, support publishing, track competitors, and measure progress in a live dashboard.

+ Everything in Fix
+ Human-reviewed re-runs on demand
+ Live visibility dashboard for work in progress
+ Publishing and outreach workflows — managed end-to-end
+ Scoped mention/citation targets; delta reviewed on re-runs
+ Schema hygiene + content PRs to your repo / CMS
+ Slack with a real goblin · <24h SLA

Start the engine →

by application

Goblin Forge

Want it fully managed by goblins? Custom build.

// multi-brand teams, scoped first

$2,000/ mo+

For teams with bigger sites, multiple domains, custom workflows, or white-label needs: a deeper managed program built around your business. Starts at $2k/mo; scoped to your stack.

+ Everything in Mention Engine
+ Consulting + engineering review on your AI-visibility stack
+ Custom agent and workflow design for your systems
+ Multi-brand scope by application — scoped at kickoff
+ Managed reporting across brand, marketing, and analytics
+ Human-approved content and outreach queues
+ White-label deliverables + quarterly strategy session
+ Limited queue; scoped operating plan before work begins

Apply →

✓ Cancel anytime. If the work isn't up to standard, we refund it — the guarantee covers the work we do, never a citation number we can't promise.

what happens after you pay

01
You pay
Stripe checkout, cancel anytime — no contract lock-in.
02
A human reviews your domain
A real goblin looks at your site and scopes the first scan — not a black box.
03
Your first scan runs
We check what ChatGPT, Claude, Gemini, and Perplexity say about you today.
04
First report lands
Within 1–2 business days — human-reviewed, not auto-fired.

Questions before you start? [email protected] · Not sure what you'd get? See a sample before/after report →

▸$ goblin --summonsolo shop · onboarding by hand

Want the scan in your inbox? Drop your domain, email, and what you want to get cited for. I'll send back the gaps and the next useful moves. Best when you need a focused scan and a concrete next move.
No card, no sales call.

$ mail[email protected]

$ chat@promptgoblin

$ ls ./elsewherelinkedin · x · github

$ pwdChicago, IL · by appt

FULLY AUTOMATED SEO AND AI VISIBILITY.FULLY AUTOMATED SEO AND AI VISIBILITY.