AI Search Optimization

Find your AI search issues, then raise a PR that fixes them, in minutes.

See how ready your site is for ChatGPT, Perplexity, and Google AI Overviews — a full GEO & AEO checkup with the evidence behind every issue, no code required.

Launching soon. Be the first to run a checkup when it's live.

  • 23 checks · 7 categories
  • Fixes in a reviewable PR
  • Zero data retention

How it works

From invisible to cited in three steps

No code to write, no consultant to hire. A checkup, a clear diagnosis, and a pull request you control.

  1. 01

    Run the free checkup

    Paste your URL. We fetch your site the way an AI crawler does and score it across 23 checks in 7 categories: rendering, structured data, metadata, crawl surface, semantics, content, and answerability.

  2. 02

    See exactly what's broken

    Get a 0–100 readiness score with per-category subscores and the precise evidence behind every check (the offending route, tag, or missing markup), plus what the agent can fix automatically.

  3. 03

    Merge the fix PR

    The agent clones the one repo you pick into an ephemeral sandbox, applies the structural and content fixes, verifies the build and types, and opens a pull request. You review and merge, and nothing ships without you.

The rubric

23 checks, grouped into 7 categories

One transparent, versioned rubric powers the free checkup, the agent's fix targets, and the post-merge re-check, so the score you're sold is the score we re-measure.

  • Rendering

    3 checks

    Is your content in the HTML before JavaScript runs? AI crawlers read the raw response, so server-rendered content, a valid doctype, and a declared charset decide whether they see anything at all.

  • Structured data

    1 check

    Valid JSON-LD (Organization and WebSite site-wide, Article and BreadcrumbList where they belong) so engines know what each page actually is.

  • Metadata

    6 checks

    Titles and descriptions sized for the SERP, self-referential canonicals, complete Open Graph and Twitter cards, favicons, and a social image that unfurls.

  • Crawl surface

    4 checks

    A robots.txt that welcomes GPTBot, ClaudeBot, and PerplexityBot, a valid sitemap, an /llms.txt index, and pages that stay eligible to index.

  • Semantics

    3 checks

    One H1, a clean heading hierarchy, page landmarks, accurate alt text, and an accessible name on every control: the machine-eye view of your page.

  • Content

    4 checks

    Descriptive internal links, visible dates and authorship, citations to trusted sources, and a clean Markdown twin of every page: the signals that correlate with getting quoted.

  • Answerability

    2 checks

    Question-shaped headings, answer-first writing, and defined terms an AI engine can lift straight into a response.

  • See where your site stands on all 23

    Run the free checkup and get the full breakdown, with the exact evidence behind every check, in seconds.

    Run free checkup

The report

A score, then the receipts

Every check returns a status, the evidence behind it (the offending route, tag, or missing markup), and whether the agent can fix it. No black box.

  • 0–100 overall score with per-category subscores
  • Pass, partial, or fail on all 23 checks
  • Reproducible: the same input always scores the same

The fix

Fixes arrive as a pull request

The agent only touches the checks it flagged, verifies the build and types in a sandbox, and opens a PR you review. Structural and faithful-reformat fixes ship automatically; net-new content stays gated behind your approval.

  • Bounded to flagged checks, never free-roams your repo
  • Build and type-checked before the PR opens
  • You review and merge, nothing ships without you

What the research shows

The signals AI engines actually reward

AI Search Optimization isn't guesswork. Independent citation research points to a consistent set of on-site signals, the same ones our rubric measures.

Answer-first

Across citation studies, the strongest predictor of whether an AI engine quotes a page was direct answerability: saying what something is in the first sentence. We score and surface it.

Cite your sources

Linking to trusted external sources (research, primary data, standards bodies) ranked among the top predictors of being quoted. We flag claims that name a source but link nothing.

Educational wins

Informational and how-to pages get quoted several times more than transactional ones. We're honest about that ceiling instead of promising citations a pricing page can't earn.

Security & trust

We touch one repo, in a sandbox, then disappear

The free checkup runs on your public pages. The fix agent runs on one repo, in a sandbox, then disappears.

Your code is never kept

We clone your repository into an ephemeral sandbox, make the fixes, open the pull request, and destroy the sandbox. Nothing persists after the run.

Only the one repo you pick is touched

Least-privilege by design. We request access to a single repository, the one you choose, and never the rest of your account or organization.

No confidential data leaves to third parties

Your source stays inside the run. We don't sell it, share it, or pass it to third-party services beyond what's needed to open your pull request.

Zero data retention, no model training

Your code is never used to train models and is not retained after the sandbox is destroyed. Readiness is measured, the fix is shipped, and nothing is stored.

Read the full security details

Find out what AI search sees when it looks at your site

The checkup is free and takes seconds. No signup, no code, just paste your URL and see your score.

Run free checkup

FAQ

Frequently asked questions

What is AI Search Optimization?

AI Search Optimization is the practice of making a website easy for AI answer engines like ChatGPT, Perplexity, Claude, and Google AI Overviews to read, understand, and cite. It extends traditional SEO with server-rendered content, structured data, an llms.txt index, and answer-first writing so engines can quote you accurately.

How is this different from traditional SEO?

Traditional SEO optimizes for ranking blue links; AI Search Optimization optimizes for being read and cited inside an AI answer. The two overlap on fundamentals like clean HTML and metadata, but AI engines also reward machine-readable structure, explicit definitions, and content they can lift directly into a response.

Is the checkup really free?

Yes. The checkup is free and needs no signup. It scores your site across all 23 checks and shows you exactly what an AI engine sees. You only pay if you want the agent to open fix pull requests for you on an ongoing basis.

Do you store or train on my code?

No. We clone the one repository you pick into an ephemeral sandbox, make the fixes, open a pull request, and destroy the sandbox. Your code is never retained and is never used to train models. Zero data retention, no model training.

Which AI engines does this help with?

The same fundamentals improve readiness across ChatGPT, Perplexity, Claude, Microsoft Copilot, and Google AI Overviews, because they all crawl the raw HTML, parse structured data, and prefer answer-first content. We optimize the signals every major engine shares rather than gaming any single one.

Can you guarantee my site will get cited by AI?

No, and anyone who promises that is misleading you. We measure and improve technical readiness, the on-site signals that make citation possible. Whether an engine cites you also depends on your content's topic and authority, which is why we're honest that educational pages get quoted far more often than transactional ones.