Inbound AI agent monitoring

Is your website ready for AI agents?

AgentSpeed grades how well ChatGPT, Perplexity, Claude, and Operator can read, navigate, and transact on your site. Free public score. Continuous monitoring from $19/mo.

No signup. Permanent public report at agentspeed.com/score/<your-domain>.

Live example

stripe.comrubric 2026.05.1us-east

checked 2026-04-23 14:30 UTC

grade91/100

discoverability92

readability88

structured_data95

actionability87

performance96

composite, last 14 scans

Checks we run

Thirty checks across five sub-scores — including MCP Server Card, OAuth Protected Resource, Web Bot Auth, Markdown content negotiation, Agent Skills, and WebMCP. Weighted, versioned, public. Failing checks emit a remediation page indexed at /score/<domain>/fix/<check>.

robots_txt.allows_known_agents

llms_txt.present

sitemap.present

markup.canonical

content.text_ratio

content.js_required_for_primary_content

content.headings_hierarchy

auth.cookie_consent_blocks_content

structured_data.jsonld_present

structured_data.validates

structured_data.type_coverage

auth.paywall_or_login_wall_on_landing

journey.primary_probe

reachability.status

performance.ttfb_ms

performance.full_render_ms

See the full rubric, weights, and grade bands at /rubric.

How it works

Submit a URL

Anonymous, no signup. We fetch the page, parse its discoverability surface, and render it once with a real headless browser.

Run the rubric

Thirty static checks plus one read-only LLM probe. Weighted into five sub-scores and a composite 0–100.

Publish a score page

A permanent, indexable page at /score/<your-domain>. Share the link; subscribe to monitor it for regressions.