Get Your Site Found.
By Google. By AI. By customers.
You have a website. Almost nobody's finding it. Type it in below and in about 30 seconds we'll show you exactly why — and what it takes to fix it. Free.
Seen by Google
Your site shows up in real search results again — not buried where no one looks.
Quoted by AI
When someone asks ChatGPT or Google's AI who does what you do, your name is in the answer.
Found by customers
The people already searching for what you sell can finally find you — and call.
How it works: we recycle your old site → get you written up in the news → make sure Google and AI both find you. That's it. You get found.
Why you need this
AI wants to see your site.
Give it the eyes to read you.
ChatGPT, Claude, Perplexity, and Google's AI answers increasingly depend on retrievable web sources — search indexes, live retrieval, and structured facts. And Google's own guidance is blunt · there's no special file or magic AI tag that ranks you. What matters is the foundation: a site that loads fast, renders clean HTML, states accurate facts, and lets the right crawlers in. Most small-business sites were built before any of that mattered. We rebuild that foundation · then add the machine interfaces emerging AI agents can use. We never promise citations · we make sure that when an AI looks, there's finally something it can read.
Built on the formula every answer engine respects
The acronyms, decoded
We optimize for every way people search — old and new.
You'll hear a lot of three-letter acronyms thrown around. Here's what each one actually means, in plain English.
Getting found on Google search.
Getting found inside AI answers (ChatGPT, Gemini, AI Overviews).
Being the direct answer voice assistants and answer boxes read out.
Making your whole site easy for AI to read and trust.
Making sure the AI models themselves can find, read, and cite you.
The signals that help systems identify, retrieve, and reconcile your business facts.
Ask the machine about yourself
What does AI say about your business right now?
Open ChatGPT or Google's AI and ask it about your business. One of three things usually happens: it doesn't know you exist, it describes a competitor instead, or it reads back hours and details that are flat-out wrong.
That's not bad luck — it's a signal problem. The AI is pulling from whatever it can read, and an older site often gives it almost nothing to work with. Our rebuild fixes the signals AI reads: clean structure, accurate facts, schema, and crawl access. We can't promise what any AI will say afterward — no one honestly can — but we make sure it has every accurate signal to work from.
By the numbers
Why this shift is real — with the receipts.
Independent industry figures, each linked to its source. These are the market's numbers, not ours — and figures change over time.
47.9%
of citations among ChatGPT's 10 most-cited domains were Wikipedia (7.8% of ALL ChatGPT citations)
Source: Profound · Aug 2024 – ↗up to 40%
more visibility in the original GEO experimental benchmark — not a guarantee of organic discovery, citations, or traffic
Source: GEO: Generative Engine Optimization (Aggarwal et al., KDD 2024) · 2023-11 (K ↗62.9%
of Google AI Overview citations did NOT rank in the organic top 10 (37.1% did)
Source: Ahrefs · 2026-03-02 ↗68.01%
of U.S. Google searches ended without a click (Jan–Apr 2026)
Source: SparkToro · 2026-06-08 ↗“If the AI can't read you, it can't recommend you. We rebuild your site so it can read you — and so does Google.”
Industry references last reviewed June 13, 2026 · figures are third-party and change over time · results vary.
The 69-crawler AI allowlist
We welcome every AI crawler · not just the famous ones.
ChatGPT, Claude, Perplexity, Google AI Overviews are the obvious 4. But there's also Bytespider (TikTok), Amazonbot, Bingbot, Applebot, MistralAI, Meta-ExternalFetcher, CCBot, Diffbot, YandexBot, archive.org_bot, and the entire 2026 fleet of new crawlers · TerracottaBot, NovellumAICrawl, AnchorBrowser, ProRataInc, Timpibot, and more. Each gets an explicit "yes, you can read this site" rule in robots.txt. We name all 69 AI crawlers explicitly — plus 3 content-use policy tokens (opt-out controls with no crawler traffic of their own). And the catch-all still welcomes everyone else.
See all 69 crawlers we welcome
Plus 3 content-use policy tokens — Google-Extended, Applebot-Extended, Cohere-Train — which are training/opt-out controls with no HTTP user-agent. We list them in robots.txt as policy, but they are not crawlers and never appear as traffic.
These are the exact 69 crawlers our scoring engine checks — the same list published in our robots.txt. View it live at /robots.txt · we eat our own cooking.
Voice assistant optimization
When someone asks Siri, Alexa, or Google for a business like yours — we make you answerable.
More people are talking to their phones and speakers instead of typing. We give the voice assistants the structured facts they need to read your name, hours, and services out loud — so you're an option when someone asks by voice.
What we build into your site
- LocalBusiness, opening-hours & service schema — the structured data engines and assistants pull facts from, matching what's visible on the page.
- Speakable markup, where it fits — the schema.org signal for the lines a voice assistant may read aloud. It's a beta feature; we add it where it applies and never oversell it.
- Answer-first content — direct, one-sentence answers an assistant can speak without stumbling.
- Directory presence — consistent name, address & phone across the listings each assistant trusts.
Which directory feeds which assistant
Bing Places still matters twice over — OpenAI documents that ChatGPT Search can use Bing as a search provider, so a strong Bing listing travels further than Bing alone.
We give each assistant the signals it documents — we can't control what it ultimately says. Voice platforms change their rules often; results vary.
What we actually do
Most businesses don't need a new website.
They need a better foundation.
We take an existing website and rebuild it from the ground up using modern technology, modern infrastructure, and modern best practices — while preserving the content, branding, reputation, and identity you've already worked hard to create.
The result
Visitors
experience a faster website
Search engines
experience a healthier website
AI systems
experience a clearer website
The 9 layers
What actually determines whether AI systems can find, trust, and cite you.
There is no magic AI tag — Google's own guidance says so. What exists is a stack of nine layers. The bottom ones are the web foundation every major search engine and AI system documents depending on. The upper ones are the machine interfaces and controls built for what's coming. We rebuild the foundation first, add the interfaces second, and tell you which layer every fix lives in.
The human web
Fast, clean, server-rendered HTML. Every layer above rides on this one. If your page is slow or paints in with JavaScript, machines struggle before anything else gets a chance to matter.
Classic search
Googlebot, Bingbot, canonical URLs, sitemaps with honest lastmod dates, IndexNow. Google is explicit: its AI answers ride the ordinary Search index — a page must be indexed and snippet-eligible to appear at all.
Structured facts
Schema.org JSON-LD that matches what's actually visible on the page. Ordinary, documented infrastructure that helps machines parse who you are — not a secret AI ranking lever. Google says no such schema exists.
AI search crawlers
OAI-SearchBot (ChatGPT), Claude-SearchBot, PerplexityBot. Each vendor documents that blocking its search crawler can reduce your visibility in its answers. We verify access at the edge — not just declare it in a file.
Agents & user fetches
ChatGPT-User, Claude-User, Perplexity-User — assistants fetching your site live, on a real customer's behalf. Real links, working buttons, and accessible semantics decide whether an agent can actually use you.
Machine interfaces
llms.txt, our /ai manifest, Markdown representations. Google ignores these for ranking — we say that out loud. They're interoperability for the emerging agent ecosystem, and we label them exactly that. Nothing more.
Content-use policy
Search, agent, and training access are three separate decisions — not one "block AI" switch. We document yours explicitly. (Google-Extended, for example, is a policy token, not a crawler — it never belongs in traffic stats.)
Edge enforcement
Your firewall or CDN can silently 403 a crawler your robots.txt welcomed. robots.txt is what you declare; the edge is what actually happens. Our scanner tests both and shows you the difference.
Measurement
Search Console, Bing's AI Performance citations, utm_source=chatgpt.com referrals in your analytics, verified crawler logs. Real outcomes from the engines' own reporting — not just a readiness score.
Layers 1–3 are the foundation. 4–5 are access. 6–8 are the interfaces and controls for what's next. 9 is how you know it's working. We give AI engines every signal they need — what they do with that is up to them. Results vary.
How it works
Three steps. You stay in control.
Free scan
Drop your web address. In 30 seconds we score how well the AI search engines can read your site · ChatGPT, Claude, Perplexity, Google AI Overviews. No card, no email.
$0 · always free · click to scan ↑
Recycle my site
We rebuild your site pixel-for-pixel on Cloudflare's edge · faster, secure, and readable by the AI engines. Includes the AI-visibility package and a press release. Your brand stays exactly as it is.
$997 one-time + $50/mo · maintenance included
We keep it alive
The $50/mo keeps your site hosted, secure, monitored, backed up, and found · every site stays on maintenance. Move up to Growth or Weekly for more hands-on care. Cancel anytime · we email you a clean static export so you can take it anywhere.
From $50/mo · Basic · Growth · Weekly →
You own your content, brand, and domain · forever. We do not touch your site unless you choose to recycle.
Industry case studies · public references
We don't promise. We point at the data.
Six companies who already migrated off slow legacy stacks and published their numbers. None are our customers · yet. They're the documented receipts that prove the playbook works · in DTC e-commerce, publishing, enterprise software, and independent consulting. Tap any one for the full story.
DTC e-commerce · public reference
Panda Patches
Panda Patches is a kids' adhesive bandage brand that got attention on Shark Tank and tried to scale a WordPress storefront. The site was loading in 4+ seconds, Lighthouse Performance score was 42 out of 100, and the founder watched conversion rate collapse on mobile. They migrated to Next.js on Vercel's edge platform · same products, same brand, faster code. Load time dropped under 1 second. Lighthouse hit 90+. Monthly revenue went from $12K to $38K — a 3.2x lift on the same traffic. The lesson: speed is conversion. The platform is the product. WordPress was eating their business.
Before
WordPress · 4s+ load · Lighthouse 42
After
Next.js on Vercel · under 1s · Lighthouse 90+ · revenue $12K → $38K/mo
Source
Founder-reported case · Vercel Community post + company site (first-party numbers)
Why we cite it
A founder-reported example of a static-edge rebuild coinciding with a ~3x revenue lift on the same traffic (first-party numbers · results vary).
Independent consultant · public reference
Kashif Aziz
Kashif Aziz is a web consultant who runs his own portfolio + blog on managed WordPress hosting, with a Lighthouse score floating in the low-70s no matter how many performance plugins he installed. He rebuilt the entire site on Astro and deployed to Cloudflare Pages, and publicly documented hitting Lighthouse 100/100/100/100 across Performance, Accessibility, Best Practices, and SEO on his rebuild (self-published first-party account). The Astro community picked the write-up as a reference. Past performance of a third-party migration; your individual results will vary by stack starting state and content complexity.
Before
WordPress · managed hosting · Lighthouse 70s
After
Astro + Cloudflare Pages · published Lighthouse 100s (source: Kashif Aziz blog)
Source
Kashif Aziz blog · Astro community feature
Why we cite it
Same stack we use (Astro + Cloudflare). Demonstrates the stack can hit Lighthouse 100s on a real consultant portfolio (source: Kashif Aziz blog · publicly documented case).
Global book publisher · public reference
Pan Macmillan
Pan Macmillan is one of the world's largest book publishers — Penguin Random House's sister, $1B+ revenue, thousands of book pages with deep metadata requirements. They were on a sprawling WordPress install that couldn't scale and got slower with every new title. They moved to a Jamstack architecture on Netlify with static-edge delivery. Page-load speed improved 8x. Critically, click-through to retailer pages (Amazon, Barnes & Noble, indie bookstores) jumped 25% — the actual conversion metric for a publisher. The migration preserved every existing URL via redirect map. No SEO loss.
Before
WordPress · slow · siloed
After
Jamstack on Netlify · 8x faster · +25% retailer clicks
Source
Netlify published enterprise case study
Why we cite it
Proves the migration works at scale (thousands of pages) without SEO regression.
Enterprise HR software · public reference
HireRoad
HireRoad is an enterprise HR SaaS that grew by acquiring three smaller competitors. They ended up with three separate domains each ranking on different keywords — and they wanted to consolidate to one domain without losing the combined SEO value. This is the migration scenario most agencies refuse to touch because the downside (a botched redirect map nukes years of ranking) is catastrophic. HireRoad executed a careful redirect map covering every URL across all three domains. The result: zero traffic loss in the immediate post-migration period, and one year later actual traffic ran 14.5% ABOVE the analysts' post-migration forecast — never dipping below anticipated traffic (vendor case study, their own numbers).
Before
3 domains · pre-merge SEO risk
After
Single domain · zero traffic loss · 14.5% ABOVE the post-migration forecast one year later
Source
HireRoad published case study
Why we cite it
Proves the 301 redirect map approach works for high-stakes domain consolidations. This is what gets included in every recycle.
News publisher · public reference
PinkNews
PinkNews is the largest English-language LGBTQ+ news outlet, with millions of monthly readers and tens of thousands of indexed pages. They replatformed off a legacy publisher CMS onto a modern Jamstack with proper schema, faster delivery, and a redirect map that preserved their existing rankings. Result: keyword rankings grew from 669,000 to 958,000, and monthly organic traffic went from 539K to 1.2M visitors — more than double (percentages derived from the case's raw numbers; agency case study by Suso Digital).
Before
669K keyword rankings · 539K monthly traffic
After
Rankings 669K → 958K · monthly organic 539K → 1.2M (derived: more than double)
Source
Ahrefs case studies database
Why we cite it
Proves that a replatform compounds — not just preserves — your organic search visibility.
Web dev publication · public reference
Smashing Magazine
Smashing Magazine is the global publication of record for front-end developers — hundreds of thousands of pages, decades of archived content, an extremely SEO-sensitive audience. They moved off WordPress to a Hugo static-site stack on Netlify. Time to first load — the metric named in Netlify's customer case — went from 800ms to 80ms, exactly 10x faster. They published the full engineering postmortem on their own blog, with code, infrastructure decisions, and trade-offs documented (vendor case study for the hosting numbers).
Before
WordPress · 800ms time to first load
After
Hugo Jamstack · 80ms time to first load (10x faster)
Source
Smashing Magazine engineering blog
Why we cite it
Proves the static-edge approach works for the most technical, SEO-sensitive audience in the world — and gives us the exact playbook.
Pricing
$997 to recycle. $50/mo to keep it found.
One rebuild, one simple monthly plan. You keep your content, your brand, and your domain — we run the build system that keeps Google and AI finding you. If we ever part ways, we email you a clean copy of your site to take anywhere.
Marina · AI Secretary
Optional monthly add-on
Too busy to answer the phone?
Under a tow rig. Up on a roof. Hands in a customer's mouth. Driving to the next job. A missed call can represent hundreds or thousands of dollars depending on your average job · and many callers simply try the next business when they reach voicemail.
Marina takes the call so you don't have to.
Set the ring delay to 1 ring or 5 · your call. When you don't pick up, Marina answers in a voice of your choice · pick from our library of professional voices, or train her on your own voice if you'd rather she sound exactly like you. She handles the same boring questions you answer every day · hours, pricing, services, are you open · and books the actual job. You get a text with the lead, the audio, and a transcript before the customer's even off the line.
No more voicemail dead-ends. No more "I'll call them back tomorrow" leads that go cold. Let your assistant do the work while you're working.
Affiliate program
10% per sale · zero residuals
You know a small business with a slow site. We rebuild it. You get 10% on every sale your link drives · roughly $100 on a recycle · plus 10% of the first maintenance or Marina invoice. One-time. Clean.
See how the affiliate program worksFor context: small-business redesigns start at $3,000-15,000. AEO consultant retainers run $3,000-25,000 a month. We sit dead-center of a vacant lane.
The white space
AI builders erase your brand. AEO consultants don't ship code.
We do both. We keep your brand intact.
| What you get |
AI builders
10Web · Durable · Wix · Framer |
AEO consultants
Avenue Z · NoGood · iPullRank | Website Recycling |
|---|---|---|---|
| Pixel-faithful mirror of your existing site | Usually generates a new template site | No · audit only | Yes |
| Brand identity preserved | No · rewrites tagline | N/A · no build | Yes · locked |
| Published auditable AI-visibility score | No | Vendor-opaque | 20-point · audited |
| Done-for-you delivery | Self-serve | Often consulting-led · no production build | Yes · shipped live |
| You own your content + domain + customer data | Export/ownership varies by platform | N/A | Yes · forever · clean static export on exit |
| Pricing model | $10-100/mo forever | $3K-25K/mo retainer | $997 once + $50/mo · maintenance included |
Why Website Recycling exists
You shouldn't have to rebuild your online presence every few years.
Preserve what works. Modernize what doesn't. Keep growing without starting over.
That's exactly what we do — rebuild your site for the modern internet and continuously prepare it for what's next.
You keep your business. We rebuild the engine.
FAQ
Real questions from real customers.
What is Website Recycling?
Website Recycling rebuilds your existing website pixel-for-pixel so artificial-intelligence search engines like ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews can understand, trust, and cite your business. We don't replace your brand · we keep it. We rebuild the bones underneath so the machines stop tripping over them.
What if I don't like the rebuild?
The rebuild is pixel-faithful to the site you already have · same colors, same words, same images, same layout. If anything is off we revise until you sign off. 60-day money-back guarantee · no questions asked. We don't ship until the comparison passes.
How is Website Recycling different from traditional SEO?
Traditional SEO chases Google rankings. We make your site easy for AI assistants to read and quote. Ahrefs' March 2026 analysis found 37.1% of Google AI Overview citations also ranked in the organic top 10 — roughly 62.9% did not (that study is Google-AIO-specific; other AI platforms vary). So being cited doesn't require ranking #1 on Google. We build the fundamentals the engines actually read: clean server-rendered HTML, direct-answer-first paragraphs, question-based headings, accurate structured data, fast delivery, and verified AI-crawler access · the foundation, not a magic tag. Google's own guidance says there's no special AI markup that ranks you · the site itself has to be readable.
Do I have to change hosts?
Your domain stays at your registrar · we point the DNS at Cloudflare's edge. Most customers cut their hosting bill substantially — Cloudflare Pages has a genuinely free static tier, versus typical managed-WordPress hosting bills. Exact savings depend on your current host and plan. Your old host can stay running side-by-side until you sign off.
What if my site is on Wix, Squarespace, or Shopify?
We rebuild from your live URL · the platform underneath doesn't matter. WordPress, Wix, Squarespace, Shopify, GoDaddy, raw HTML · we crawl what AI engines see, mirror it pixel-faithfully, and ship a static version that loads in under a second. If you're on Shopify and want to keep the storefront on Shopify, we recycle the marketing pages and bridge the cart at checkout.
How long does the rebuild take?
It takes what it takes · we don't promise stopwatch times. The work moves through named stages: crawl, audit, rebuild, pixel-diff sign-off, deploy. You see the stage you're in at every step. We don't ship until the Lighthouse and schema checks pass cleanly.
What happens to my SEO?
We protect the signals you've already earned. We map every existing URL to a 301 redirect, preserve every page title and meta tag, keep author and date metadata intact, add accurate structured data, and submit a fresh sitemap to Google Search Console the day we deploy — the technical foundation a clean migration depends on. Well-executed migrations often improve performance and visibility; public examples include HireRoad landing 14.5% above its post-migration traffic forecast a year later, PinkNews keyword rankings 669K→958K, and Pan Macmillan +25% clicks to retailers (each a vendor or agency case study) — those are their outcomes, shown as examples, not a promise of yours. Google is explicit that meeting best practices does not guarantee crawling, indexing, or ranking, and no legitimate provider can promise otherwise. The redirect map is the deliverable; results vary.
Who owns the rebuilt site?
You own your content, your brand, your domain, your customer data · forever. We own the proprietary build system that powers the site (the Self-Updating Site AI, the AI-discovery integrations, the scanner, the backup tooling). Think of it like buying coffee at a coffee shop · you own the coffee, you don't own the espresso machine. Your site is a managed subscription · the $50/mo keeps it live, hosted, maintained, and found. If you ever cancel, we email you a clean static export of your site (HTML, CSS, images · no proprietary integrations) so you can host it anywhere. No platform lock-in. Your domain and content always stay yours.
Will the AI engines actually start citing me?
Honest answer · documented small-business citation case studies are still rare in 2026, the format is too new. What we control is the structure · clean static HTML, structured data, definition-first paragraphs, correct indexing, sitemaps, and verified AI-crawler access · the foundation every search engine and AI system actually reads · plus published llms.txt and an explicit AI-crawler policy for the agents that use them. That's the readable base; what the engines do with it is up to them. Profound's Alchemy customer case study reports a 7x higher signup rate from AI-referred visitors than other traffic sources (a vendor customer case, not a universal benchmark) · results vary by industry and execution.
Can I update the site myself afterward?
Yes. Every recycle ships with a one-click editor · updating hours, prices, photos, or copy is a click-and-save operation. No developer needed. If you'd rather email us "change my hours" we'll do it on the same day.
Do I have to film a proof video myself?
No. Our Marina AI agent calls you for a 90-second voice interview. You can pick a professional voice from our Cartesia Sonic 2 library, OR opt to have your own voice trained from the interview recording. Either path, HeyGen Avatar IV renders a 60-90 second talking-head video using your selected voice reading the rebuilt-site pitch. You never touch a camera.