Would an AI agent succeed with your documentation?
Your customers' coding agents might be reading your docs before they do. If an agent can't find, read or trust what it needs, the integration fails and your product takes the blame. This free checklist covers the 16 things that fix that, in the order an agent hits them.
Free, no payment, no cardBased on real-world AI agent behaviour
- A priority for every check, so you know where to start
- A step-by-step guide behind every check
- Built on open work by AFdocs, Google's OKF, and manni
Four layers, in the order agents hit them
Most teams stop after the first two. The last two are where agents give confident, wrong answers.
Find
Can it reach the right page?
- Reachable
- Stable URLs
- An llms.txt index
- Machine-readable API specs
Read
Does it get the content intact?
- Server-rendered pages
- Markdown on request
- Pages that fit the window
- Tabs that flatten cleanly
Understand
Is there enough to act on?
- Nothing left to guess
- Complete code samples
- One term per concept
- Consistent frontmatter
Trust
Is it still true today?
- Procedures that still work
- A spec that matches reality
- Tested samples, live links
- Labelled versions
Get the full checklist
16 checks, each with a priority and a step-by-step guide. Free.