Four pillars, up to 32 scored checks, and your market’s own questions put to ChatGPT, Perplexity and Gemini on a schedule — measured together, because the engines read them together.
Before AI recommends anyone, it reads the local layer: your profile, your reviews, and whether every listing agrees on who you are. This is the pillar most businesses think is handled — and the one that quietly drifts.
Read the Local SEO guide →Classic search still drives the majority of visits — and every AI engine builds its understanding of you on top of it. A slow, unindexable, thin site loses both games at once.
Read the SEO guide →Answer engines quote; they don’t browse. Structured data tells them exactly who you are, and answer-shaped content gives them something worth lifting — into featured snippets today and AI citations tomorrow.
Read the AEO guide →The layer nobody else shows you: whether AI can read your site at all, and whether it names you when your customers ask. We test crawler access from the outside and run your market’s real questions on a schedule.
Read the GEO guide →The same crawler our customers pay for, pointed at your domain right now. It fetches as a browser, Googlebot, GPTBot, ClaudeBot and PerplexityBot, samples up to 25 pages, and grades them across Local, SEO, AEO and GEO. Usually under a minute.
Three free scans a day. No email, no card.
Create an account and scan up to 5,000 pages, on a schedule, with the whole fix list and every check — not just the top three.
Create your account → Free scans reset 24 hours after your first one.Each pillar is scored 0–100 from its checks, then weighted into one Speak Local Score — so a firewall blocking GPTBot hurts, a five-star review surge helps, and your team always knows why the number moved.
Profile, reviews, NAP, local pack — the trust layer.
Technical health, rankings, content, authority.
Schema coverage and quotable, answer-shaped content.
Crawler access, entity clarity, AI citation share.
Enter your domain and we crawl it the way Google, GPTBot, ClaudeBot and PerplexityBot do — up to 10,000 pages, seeded from your sitemap. Then add the questions your market actually asks, and we put them to the assistants on a weekly or monthly round.
One number across Local, SEO, AEO and GEO for your own site. Then a separate standing against five named rivals — link authority, organic reach, search visibility, AI answers and site health — built the same way for every domain.
Work a prioritized, plain-English fix list: schema to add, crawler blocks to lift, content to reshape — each with the why and the how.
Scheduled re-scans, with a notification when a passing check starts failing, when a fix is verified fixed by an independent re-check, when new reviews land — and, the part most tools stay quiet about, every time scheduled work was skipped and why.
One script tag and ChatGPT, Perplexity, Claude, Gemini, Copilot and a dozen more each become a named source instead of being buried in Referral. We also show arrivals whose source was stripped rather than quietly filing them as Direct. What it can’t see, and the screen says so: a JavaScript tag never sees GPTBot, because crawlers don’t run JavaScript.
Clicks, scroll depth and hesitation drawn over an actual screenshot of your page, captured at phone, tablet and desktop width with real touch emulation — so a phone click lands on the phone rendering. Then replay the sessions that went wrong, with every form field masked in the visitor’s browser before anything reaches us.
Point Claude at your own measurements and ask which clients slipped this week — twenty read-only tools over MCP, with an API key you can revoke in a second. Read-only is structural, not a promise: the tools module cannot import a vendor client at all, so no prompt can start a scan or spend your allowance.
Your site and every rival scored 0–100 the same way — link authority, organic reach, visibility on your own tracked searches, AI answers, and site health.
A signal we can’t evidence is left out and its weight shared among the rest — drawn as a dash, never a zero, and the page says which. Every scale is absolute: with three or four domains, stretching them to fit turns a three-point gap into a chasm. Inside five points it reads level, never a lead. And site health alone never produces a rank.
Audit, score, fix and monitor how you show up in Google and in AI answers.