Four pillars, up to 32 scored checks, and your market’s own questions put to ChatGPT, Perplexity and Gemini on a schedule. We measure them together because the engines read them together.
Before AI recommends anyone, it reads the local layer: your profile, your reviews, and whether your listings agree on who you are. Most businesses think this part is handled, and it’s the one that drifts without anyone noticing.
Read the Local SEO guide →Classic search still drives most visits, and AI engines build their picture of you on top of it. A slow, thin site that can’t be indexed loses on both.
Read the SEO guide →Answer engines quote short passages from your pages. Structured data tells them exactly who you are, and answer-shaped content gives them something worth quoting, in featured snippets and in AI citations.
Read the AEO guide →This pillar covers what other tools leave out: 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 →Run the same crawler our customers pay for on your domain 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. It usually takes under a minute.
Three free scans a day. No email, no card.
Each pillar is scored 0–100 from its checks, then weighted into one Speak Local Score. A firewall blocking GPTBot pulls it down, a run of five-star reviews pushes it up, and your team can always see why the number moved.
Profile, reviews, NAP and local pack.
Technical health, rankings, content, authority.
Schema coverage and quotable, answer-shaped content.
Crawler access, entity clarity, AI citation share.
What each pillar means for a local business: AI local SEO · AEO for local businesses · GEO for local businesses · local AI visibility
Enter your domain and we crawl it the way Google, GPTBot, ClaudeBot and PerplexityBot do: up to 10,000 pages, starting from your sitemap. Then add the questions people in your market ask, and we put them to the assistants weekly or monthly.
Your own site gets one number across Local, SEO, AEO and GEO. You also get a separate standing against five named rivals, built from link authority, organic reach, search visibility, AI answers and site health, and calculated the same way for every domain.
Work through a prioritized fix list: schema to add, crawler blocks to lift, content to reshape. Each item says why it matters and how to do it.
Scheduled re-scans notify you when a passing check starts failing, when an independent re-check confirms a fix, and when new reviews land. Unlike most tools, Speak Local also tells you every time scheduled work was skipped, and why.
Add one script tag and ChatGPT, Perplexity, Claude, Gemini, Copilot and a dozen more each show up as a named source instead of being buried in Referral. We also show visits whose source was stripped, instead of filing them as Direct. The screen also says what it can’t see: a JavaScript tag never sees GPTBot, because crawlers don’t run JavaScript.
Clicks, scroll depth and hesitation drawn over a real screenshot of your page, captured at phone, tablet and desktop width with real touch emulation, so a phone click lands on the phone layout. You can also replay the sessions that went wrong. Every form field is masked in the visitor’s browser before anything reaches us.
Point Claude at your own measurements and ask which clients slipped this week. It gets twenty read-only tools over MCP, with an API key you can revoke in a second. The read-only part is enforced in code: the tools module cannot import a vendor client at all, so no prompt can start a scan or spend your allowance.
Your site and each rival are scored 0–100 the same way, from link authority, organic reach, visibility on your own tracked searches, AI answers and site health.
If we can’t evidence a signal, we leave it out, share its weight among the rest and draw it as a dash rather than a zero, and the page says which. Every scale is absolute, because with three or four domains, stretching the scale to fit turns a three-point gap into a chasm. A gap under five points reads as level, not a lead. Site health alone never produces a rank.
Audit, score, fix and monitor how you show up in Google and in AI answers.