Free tool · No signup

Can an AI agent read your docs?

Your documentation now has two audiences: the developers who evaluate your platform, and the AI agents they send ahead. This scan fetches your public docs the way an agent would and runs 20 checks from the DevEx Friction Audit framework. Results in about 30 seconds, with the evidence for every finding.

Free, no signup. We fetch your public pages once, score them, and store nothing.

or try

What gets checked

Four categories, drawn from the audit framework

The scan samples the dimensions of the DevEx Friction Audit that are visible from outside, with no account and no credentials. Agent readiness comes first because the conventions there are young enough that most platforms haven't decided what to do about them, which makes it the cheapest gap to close and the easiest place to get ahead.

01

Agent readiness

llms.txt, markdown for machines, a fetchable OpenAPI spec, MCP surface, JavaScript-free readability, AI crawler access. Scored against the same AU1 to AU10 rubric the paid audit uses.

02

Getting started

Whether a new developer can find the quickstart, copy runnable code, make a curl request, and see the path to credentials without leaving your docs.

03

Documentation quality

Search, changelog, error reference, rate limit docs, and freshness signals. The pages integrators need at 2am, checked for existence and discoverability.

04

Structure and metadata

Heading hierarchy, page weight in tokens, canonical URLs and social metadata. What your docs cost a machine to read and index.

What this is, and is not

This is a surface scan: about a dozen HTTP requests against your public pages, scored by heuristics and stored nowhere. It cannot execute your quickstart, judge your prose, or explain why an agent abandoned a task halfway through, so a strong result means the basics are in place rather than that the work is done. A weak one is the cheapest warning you will get this quarter.