Agent readiness checkers compared
An agent readiness checker fetches your site the way an AI agent would and scores how well agents can reach it, read it and use it. The main free ones as of November 2026 are Vercel's Is Agentic, Cloudflare's isitagentready.com, Frase, Backlinko, SEOJuice and AgentsReady (this site), plus Mintlify's and Fern's scores for developer docs. They agree on the basics (no bot wall, server-rendered content, sensible robots.txt) and differ on what else they weigh: emerging protocols, whether ChatGPT cites you, docs quality, or whether an agent can finish a job.
From our data: How AgentsReady scores: 60 core points across 14 checks every site needs, plus 40 points for one job (be recommended, buy, sign up or get an API key). As of 3 Nov 2026 we rank 214 software vendors across 20 categories with the same scan.
The checkers
Is Agentic (Vercel, scans by Ora)
0 to 100 from Essential checks (server rendering, HTTP behaviour, structure, usable controls, identity metadata) plus Recommended checks that only apply when it finds an API, OAuth, MCP or commerce surface. Shows evidence and a fix per issue, and an observed agent journey that doesn't count toward the score. Free, with a CLI (npx is-agentic), a JSON API and a read-only MCP server.
isitagentready.com (Cloudflare)
Launched 17 Apr 2026. The widest coverage of new standards: Link headers, markdown negotiation, Content Signals, Web Bot Auth, MCP server cards, Agent Skills, WebMCP, API Catalog, OAuth discovery, plus commerce protocols checked but not scored. Gives a prompt for a coding agent per failing check. Free.
Frase Agent Readiness
Access, discovery and parseability checks, 0 to 100 with clear bands, and a null score when it can't grade fairly. Free to run; the assembled report with paste-ready fixes needs a free trial of Frase.
Backlinko AI Agent Readiness (Semrush)
Splits the result into Agent Access and AI Citation, and runs 15 ChatGPT prompts to show whether you're mentioned and who is named instead. Free, three checks a day.
SEOJuice Agent Ready
Weights API, auth, MCP and skill discovery at 45%. Reports show raw HTTP evidence, a copyable developer brief, score history and a benchmark against similar sites. Free, five scans a day.
Mintlify Agent Score and Fern Agent Score
Docs-only scores with letter grades and public leaderboards. Fern's is built on the open AFDocs spec (npx afdocs check).
AgentsReady (this site)
Free. Scores the job, not just the page: a store is checked on product data and whether the cart opens for agents, a SaaS product on pricing and sign-up, a developer platform on OpenAPI, MCP and self-serve keys. Can send a real agent to try the job, and ranks vendors by category from the same scan.
Which one to use
A content or marketing site
Frase or Backlinko for AI visibility and citations; AgentsReady with the Be recommended job for crawler access, contact details, Organization and FAQ data.
An online store
AgentsReady with the Buy job: product and offer data, prices in text, and whether /cart opens for an AI user-agent. isitagentready also lists UCP and ACP support.
A SaaS product
AgentsReady with the Sign up job for pricing, sign-up reachability and free-plan signals; Is Agentic for a broad technical review.
Developer docs and APIs
Mintlify, Fern or afdocs for docs quality; AgentsReady with the Get an API key job for OpenAPI, MCP, OAuth discovery and keys.
Tracking new standards
isitagentready.com. It adds checks fastest, and Cloudflare publishes adoption data for each standard in Radar.
Why the scores differ
Each tool picks different checks and weights them differently, and some exclude checks that don't apply to a site while others count them as failures. The same docs site can score in the 80s on one tool and much lower on another. Compare a tool's score with itself over time, read the failing checks rather than the number, and fix the ones that stop an agent first: bot walls, missing server-rendered content, and a blocked action path.
Where AgentsReady is weaker
We don't yet score Content Signals, Web Bot Auth, Agent Skills indexes, API catalogs or WebMCP, which isitagentready checks. We don't ask ChatGPT whether it cites you, which Backlinko does. And our docs checks are lighter than AFDocs'. If those matter to you, run those tools alongside ours.
Check it yourself
The first two commands are the test every checker starts with: does an AI user-agent get your page? The others run Vercel's and the AFDocs checks from your terminal. For the full job-based score, run the free check here.
curl -s -o /dev/null -w "%{http_code}\n" \
-A "Mozilla/5.0 (compatible; Claude-User/1.0)" https://example.com/
curl -s -o /dev/null -w "%{http_code}\n" \
-A "Mozilla/5.0 (compatible; ChatGPT-User/1.0; +https://openai.com/bot)" https://example.com/
npx is-agentic example.com --json
npx afdocs check https://docs.example.comQuestions
Which agent readiness checker is the most accurate?
None measures everything. They agree on the basics, so if two tools flag the same failure, fix it. For whether an agent can actually finish a task, the most direct evidence is a real agent run, which AgentsReady and Is Agentic both show.
Are agent readiness checkers free?
All the ones on this page can be run free. Some cap daily scans (Backlinko three, SEOJuice five) and some keep the full report behind a trial (Frase). AgentsReady has no paid plans.
Why did my site score differently on two checkers?
Different checks, different weights, and different rules for checks that don't apply. Read the failing checks instead of comparing numbers across tools.
Should I chase a perfect score?
No. Fix what blocks agents first: bot walls, content that needs JavaScript, and a gated action path. Bonus checks for brand-new standards matter less until agents use them.