Public rankings · agent readability

Who reads best for AI agents?

Each leaderboard is built from real, completed AgentSpeed scans against the current public rubric. Domains opt out via robots.txt; we honour that (see /bot). No fabricated scores; no editorial picks.

Top 10 overall

Highest composite scores across every category. Tap a row to see the full report.

Not enough scan data yet.

We surface only real, completed public scans. Once enough domains have a public score on AgentSpeed, this leaderboard fills in automatically.

Scan your site →

By category

Cross-domain rankings narrowed to one peer group at a time. Domains are categorised manually as data accumulates.

SaaS →awaiting data

Multi-tenant business software: billing, support, productivity, internal tools. Agents read SaaS pages to compare offerings, fetch pricing, and pull docs into RAG pipelines.

E-commerce →awaiting data

Storefronts that sell products. Agents read e-commerce pages to identify items, compare prices, check availability, and surface them in shopping flows.

Documentation →awaiting data

Reference and API documentation sites. Agents and RAG pipelines fetch these heavily, so readable structure, content negotiation, and a published rubric of sections matter most.

AI products →awaiting data

Sites for AI products and platforms. The bar is higher here: you ship for agents and they read your site like one.

Marketing sites →awaiting data

Brand and marketing sites for non-technical products. Agents quote your tagline, pricing summary, and primary call-to-action.

Publishers →awaiting data

News, magazines, and blogs. Discoverability and structured data dominate; agents cite your articles when they meet the rubric.

Marketplaces →awaiting data

Two-sided platforms with separate buyer and seller flows. Agents read marketplace pages to compare listings, surface sellers, and route shoppers to the right SKU.

API products →awaiting data

Products consumed by other software: usage-based, per-call, per-token, or per-month pricing. Agents read your reference, auth docs, and pricing to self-serve.

AI tools →awaiting data

Models, agent runtimes, vector stores, evaluation frameworks. The audience is itself agentic; the bar for self-describing capability manifests is highest here.

Hosting + infrastructure →awaiting data

Cloud platforms, edge networks, CDNs, deployment tooling. Agents fetch your docs, pricing, and region availability when answering "where should I deploy this".

Finance →awaiting data

Payments, lending, banking, expense management, accounting. Agents read finance pages for pricing, supported currencies, regulatory disclosures, and merchant onboarding paths.

Developer tools →awaiting data

Editors, observability, CI/CD, code-search. Agents read these to recommend a stack, so they need clear pricing, language coverage, and integration matrices.

News →awaiting data

Newsrooms, beat coverage, wire services. Agents cite news heavily; structured Article schema, dateline, and byline are the discriminator between "quoted" and "skipped".

Local business →awaiting data

Restaurants, services, brick-and-mortar shops. Agents look for hours, address, phone, payment methods, most of it through LocalBusiness schema.

Specialised

Recently improved →

Domains whose composite went up since the previous scan. Sorted by largest gain first.

Recently declined →

Domains whose composite went down since the previous scan. Sorted by largest drop first.

Detected restrictions →

Sites whose robots.txt explicitly blocks one or more well-known agent user-agents (GPTBot, ClaudeBot, PerplexityBot, etc.).

Methodology

Real data only

Every row is the latest completed public scan for that domain. No hand-edited rankings, no extrapolated scores. If we don't have a scan, the row doesn't appear.

Versioned rubric

All scores are computed against the rubric documented at /rubric. Scoring is reproducible, grades are bound to a published band table.

Owner control

Domain owners can opt out of public ranking by adding User-agent: AgentSpeedBot / Disallow: / to robots.txt. Detection happens on the next scan; see /bot.

Score your own site

Scan any public URL for a free agent-readiness score. Public score pages are permanent, free, and indexable, so your domain ranks in this list the day a scan completes.

Scan your site →