Content Discoverability
Category scoring, check messages, and fix suggestions aligned to the same flow as the original Agent Score.
Is your documentation ready for AI agents?
When an agent cannot use your docs, the user does not get an error. They simply switch to another stack. Paste a documentation URL to see the score, risk signals, and AFDocs recommendations.
Checks
22
Categories
7
Output
Scorecard
How it works
Markdown Availability
Category scoring, check messages, and fix suggestions aligned to the same flow as the original Agent Score.
Page Size and Truncation Risk
Category scoring, check messages, and fix suggestions aligned to the same flow as the original Agent Score.
Leaderboard
Agent score directory
Methodology
22 checks across 7 categories
Every run uses AFDocs to fetch your documentation directly, deterministically sample up to 10 links, and score the agent experience across discovery, markdown delivery, structure, freshness, and access.
Sample size
10 links
The crawler scores a deterministic slice of the documentation so repeated runs stay comparable.
Request pacing
100ms
Requests are spaced out between fetches to avoid hammering the target documentation site.
Timeout window
15s
Slow or inaccessible pages are surfaced during the same pass instead of being silently ignored.
6 checks
llms.txt and Discovery
Verifies whether agents can discover your documentation index, follow the links it exposes, and find llms.txt from the pages you publish.
2 checks
Markdown Delivery
Checks whether the site offers clean markdown through .md URLs or content negotiation instead of forcing agents through bloated HTML.
4 checks
Page Size
Measures whether pages fit within agent context windows and whether the useful content starts early enough to avoid truncation.
3 checks
Content Structure
Evaluates tabs, headers, and code fences to ensure the content remains parseable after an agent converts or serializes the page.
2 checks
URL Stability
Confirms that documentation URLs resolve cleanly, return correct status codes, and avoid redirect behavior that can confuse crawlers and agents.
3 checks
Observability
Compares freshness and parity signals so agent-facing content stays accurate across llms.txt, markdown output, and cached HTML responses.
2 checks
Authentication
Tests whether agents can access the docs without hitting auth walls, or whether alternate public paths exist when the main site is gated.