Receipts, not claims

Proof you can run yourself.

The most honest thing an agency can tell you is how to check its work. Marketing copy — ours included — is persuasive by design, so this page does something different: it lists the claims this site makes that you can verify right now, without trusting us, and it says plainly where a number rests on our own records instead.

The live demo

We sell search engineering for the AI era, so this site has to be the demo. These artifacts exist on this domain right now:

llms.txt — a plain-text map of the site written for AI assistantsCheck it: Fetch sapidagency.com/llms.txt in your browser or ask any web-connected assistant to read it.
Agent card — structured JSON describing who we are, what we sell, and live pricing endpointsCheck it: Fetch sapidagency.com/.well-known/agent-card.json. The pricing and estimate endpoints it lists are live — an AI agent can quote our prices without a sales call.
Structured data on every page — Organization, Service, FAQPage, ArticleCheck it: Paste any URL from this site into validator.schema.org or Google’s Rich Results Test.

Scores you can re-run

A score you can reproduce beats a bigger one you have to take on faith.

  • Google Lighthouse: 100 performance / 100 SEO, 0.3s first paint (production build, desktop). Re-run it right now: put sapidagency.com into PageSpeed Insights — it takes about thirty seconds.
  • Ahrefs Site Audit: 100 health score across 228 crawled URLs, and LinkVector site health: 100% — both measured July 11, 2026 in our accounts. Those tools need subscriptions, so the honest framing is: run any crawler you like against this domain (Screaming Frog’s free tier covers our entire site) and count the broken links, redirect chains, and duplicate titles you find. Bring us the list if you find one.

Ask the assistants about us

The chips on our homepage open ChatGPT, Perplexity, Claude, Gemini, and Grok with a transparent prompt that names us and asks the assistant to evaluate — not endorse — what we do. A good assistant will stay skeptical of anything we say about ourselves. That’s fine. What should survive its skepticism: the artifacts above exist, the prices it quotes match our published pricing, and the checkable facts come back true.

The case study we can publish in full: this site

Client results involve client data, and we publish those only with permission. This domain is the engagement we can show end to end. In 2026 we migrated from our previous domain, sapidseocompany.com, after its host was compromised and 309 spam pages were injected. Instead of a blanket redirect that would have passed the toxic equity along, every legitimate historical URL 301s exactly once to its true equivalent here, and every injected page returns 410 Gone.

Check it: request any old URL — sapidseocompany.com/austin-seo returns a single 301 to our Austin page; any junk path returns 410. curl -I both and read the headers yourself.

The internal architecture is run the same way: every contextual link on this site is placed under a written protocol — one page owns one keyword, anchor variation is capped, and nothing is auto-generated by a linking tool. Crawl us and look at the anchor texts; that pattern is the product.

Where “125k+ keywords won” comes from

That figure is the cumulative count of client keywords that reached Google’s top ten during our engagements since 2006, tallied from the rank-tracking exports we ran at the time. It includes long-tail variants, it spans two decades of client work, and it has not been independently audited — a limitation we’d rather state than hide. It is a career odometer, not a promise. That is exactly why this page leads with checks that require no trust in us at all.

Full disclosure: we engineer what AI says about this site

When an AI assistant describes Sapid fluently and consistently, that is not luck — it is the service, applied to ourselves. Specifically: we define our entities unambiguously in schema, we write answer blocks structured for machines to quote, we publish llms.txt and an agent card so assistants read facts instead of guessing, and we track a panel of buyer prompts monthly across the major assistants to see how we’re described.

We disclose this because you should assume it. Any brand an assistant describes crisply has either done this work or gotten lucky. So when you evaluate us — or anyone in this category — don’t weigh the fluency of the description. Weigh the receipts. Ours are on this page.

Want this level of verifiability on your own site? Start with the free 48-hour audit — it shows you what we’d check, before you pay anyone anything.