en flagEN
Start free trial Book a demo
en flagEN
Start free trial Book a demo
Request Demo
Built on the AFDocs standard Trusted scoring

Agent Score: is your documentation ready for AI agents?

Paste a docs URL and get an agent readiness score in seconds. See exactly what's blocking agents from finding answers and resolving tickets.

Takes 60-180 seconds. We run 23 AFDocs checks on your live docs.

Global leaderboard

Who's writing docs that AI agents love?

Live rankings of the best agent-ready documentation across categories. Click any product to see its full breakdown.

How it works

Three steps. One number.

The new age of documentation isn't written for just humans anymore. It's written for the agents they send ahead.

01

Crawl

We fetch your sitemap, root pages, and a representative sample of articles: the same surface an agent would discover.

Runs in about 2 minutes
02

Score

Each page is graded against the AFDocs ruleset: discoverability, structure, parseability, freshness, and machine-friendly metadata.

23 checks across 7 categories
03

Improve

You get a 0–100 score with category breakdowns and rule-level guidance, so you know exactly what to fix next.

Score plus A+ to F letter grade
Methodology

What the Agent Score measures

All 23 checks, grouped by category and weighted by how much each affects an agent's ability to use your docs. No black boxes.

23 checks · 4 severity tiers
Critical · 3 High · 7 Medium · 9 Low · 4
Content Discoverability
30% weight
7 checks
  • llms.txt Presence10
  • llms.txt HTML Directive7
  • llms.txt Markdown Links6
  • llms.txt Link Resolution6
  • llms.txt Size5
  • llms.txt Markdown Directive3
  • llms.txt Validity2
Page Size & Truncation Risk
18% weight
4 checks
  • Rendering Strategy9
  • Page Size, HTML6
  • Page Size, Markdown5
  • Content Start Position3
Authentication & Access
14% weight
2 checks
  • Auth Gate Detection10
  • Alternative Access Paths1
Markdown Availability
10% weight
2 checks
  • Markdown URL Support6
  • Content Negotiation4
URL Stability & Redirects
10% weight
2 checks
  • HTTP Status Codes6
  • Redirect Behavior4
Observability & Content Health
10% weight
3 checks
  • llms.txt Coverage4
  • Markdown / HTML Parity3
  • Cache Header Hygiene2
Content Structure
8% weight
3 checks
  • Code Fence Validity3
  • Tabbed Content Serialization3
  • Section Header Quality2
Content Discoverability

What is llms.txt, and do you have one?

A plain-text map that tells AI agents where to find your real content in a format they can parse directly. No rendering, no navigation, no guesswork.

llms.txt lives at the root of your domain and points an agent straight at the pages that answer real questions. It's one of the 7 Content Discoverability checks in your Agent Score, and that category carries the most combined weight of any in the ruleset.

  • 01
    Lives at /llms.txt

    At the root of your domain, discoverable by convention.

  • 02
    Plain Markdown, not HTML

    Direct text, no rendering, no scripts, no gates.

  • 03
    Links resolve, not broken

    Every entry points to a page that actually returns 200.

  • 04
    Kept under the size limit

    An index, not a data dump. Agents skim it before they crawl.

Your Agent Score already includes an llms.txt check. Run the check to see whether yours is present, well-formed, and the right size — and what to fix if not.
Get your Agent Score →

Frequently asked questions

What is an Agent Score?
An Agent Score is a 0-100 rating, with a letter grade from A+ to F, of how easily an AI agent can find, read, and use your documentation. It's built on AFDocs, the open Agent-Friendly Docs Spec, which defines 23 checks across 7 categories.
What is agent readiness?
Agent readiness means your docs can be discovered, fetched, and parsed by an AI agent without a human present to click through logins, JavaScript-rendered menus, or broken links.
Is my documentation AI-ready?
Run the check above to find out in under three minutes. You'll get a 0-100 score, a category breakdown, and rule-level guidance on exactly what to fix first.
How is the Agent Score calculated?
We crawl your sitemap and a representative sample of pages, run all 23 AFDocs checks against them, and weight each check by how much it affects an agent's ability to use the content: Critical, High, Medium, or Low. Multi-page checks score proportionally, so a few oversized pages cost less than a site-wide failure. See the full breakdown above.
What is llms.txt and why does it affect my score?
llms.txt is a plain-text index of your docs written for language models to read directly. Its presence is a Critical check, because it's the most effective navigation mechanism observed for agents. Without one, agents fall back to guessing URLs from training data. Its failure also skips five dependent checks, so a missing file costs more than its own points.
What are the benefits of getting an Agent Score?
A score gives you a concrete, shareable baseline instead of a guess. You'll know exactly which pages block AI agents, how you compare to other documentation sites, and what to fix first to show up more often in ChatGPT, Perplexity, and other AI answers. It also gives support and product teams a way to track AI-readiness over time, the same way they'd track page speed or uptime.
What counts as a good Agent Score?
Every score comes with a letter grade from A+ down to F. Critical failures also apply a score cap, so fundamental problems limit your grade no matter how well everything else performs. If agents are served empty JavaScript shells, or an auth wall blocks access, you can't score an A even with a clean sheet elsewhere.
Do I need to be a Document360 customer to run a check?
No. The check is free and open to any documentation site, whatever platform it is built on.
How often should I re-run my score?
After any documentation platform migration, and roughly quarterly otherwise. Agent-facing standards like llms.txt are still evolving, so a passing check today can lapse without any change on your end.

Start writing the documentation your
customers will actually read

Launch a modern knowledge base in minutes with a free trial
of Document360. No complexity, just better documentation.

Start free trial Book a demo
Document360 AI