A readiness scanner for the agentic web
See whether AI agents can discover, understand, and safely navigate your public website—with evidence for every finding and no mystery-grade nonsense.
Public. Read-only. No account required for the free result.
48readiness checks
19compatibility checks
0carts poked
100%of findings shown
SEO had its era
AI assistants already research, compare, and buy on behalf of your customers. If your website only works for human eyeballs — unlabeled buttons, no machine-readable structure, contradictory facts across pages — those agents bounce, and you never even know you lost the sale.
Being mentioned in a chatbot's reply — answer engine optimization — is only one small part of the picture. AgentOptimize goes further: it checks whether agents can actually use your site. Can they find your offering through your navigation? Pull out the key facts accurately? Follow your links to a clear next step? Our 67-check catalog covers discovery, comprehension, usability, and task completion across 48 readiness checks, including three A2A Agent Card checks for sites that expose an agent. The 19 agent-compatibility checks cover crawler policies, structured-data conventions, and agent protocols. All 67 appear in one profile-aware result.
One earns you a mention; the other earns you the customer.
Not another vibes-based score
AgentOptimize examines the public, machine-readable signals that help agents find your business, understand what you offer, and follow a useful journey without taking consequential actions.
The free result shows the complete diagnostic—not three alarming red flags followed by a paywall wearing a lab coat.
What it checks: your sitemap, crawler rules, and whether key pages are reachable through stable links — and whether the site looks the same however it's read.
Why you'd care: if an agent can't find your product pages, it can't recommend them. Broken sitemaps and redirect loops are invisible to you and fatal to them.
What it checks: whether your business identity, offer, prices, availability, and policies are stated plainly — and whether your visible text, structured data, and feeds all tell the same story.
Why you'd care: contradictions between pages get misquoted to your potential customers. Vague offers get skipped entirely.
What it checks: buttons with clear names, labeled forms, visible loading and error states, working search and filters — and no popup ambushing the journey.
Why you'd care: a control an agent can't understand is a control it won't press. A popup you'd dismiss in half a second can be a brick wall for an agent.
What it checks: the optional extras — feeds, APIs, catalogs, and agent protocols like MCP — are checked only where they exist, and judged on whether they're actually usable.
Why you'd care: most sites can ignore this section completely. If you've built machine-friendly doors, they should be findable and documented — otherwise they're decoration.
What it checks: that everything traces back to an accountable business — real policies, real contact paths, secure attributable links, and no secrets lying around in public.
Why you'd care: agents are cautious on behalf of their users. No accountable business, no trust — and no recommendation.
What it checks: one bounded, read-only journey: find your main offering, pull out the key facts accurately, and land on a clear, safe next step — a human, support, signup, or checkout.
Why you'd care: this is the whole point. No forms submitted, no accounts created, no cart surprises, no tiny robot buying 400 hats.
Selling on Shopify? Agent-assisted checkout, yes: agents can prepare a checkout that the shopper confirms and pays for. Fully autonomous checkout, no: Shopify's own policy blocks it, so we never count it against your store. Your part is keeping prices, availability, shipping, and returns consistent everywhere.
Plus 19 compatibility checks
Bot diplomacy. These 19 checks cover crawler policies, AI-crawler rules, sitemaps, content signals, API and auth handshakes, agent skill listings, and commerce declarations. They appear with the readiness checks in one profile-aware score; neutral and optional entries stay visible without affecting it.
agentoptimize.app / public readiness scan
We’re finishing the boring-but-important parts: safe sampling, truthful inconclusive states, private results, and a complete 48 + 19 check report.
No waitlist form yet. Your inbox has suffered enough.