Documentation

Measure whether your website is visible, understandable and usable by AI agents — then, if you sell online, whether it is also purchasable — from the dashboard, API or command line.

canagentsbuyfromyou evaluates a website from the point of view of an AI agent. You provide a URL; you get the elements that help an agent find, understand and use that site, along with recommended fixes. For an online store, the analysis can additionally check the catalog and buying journey.

These docs cover the authenticated dashboard and integrations over the HTTP API, from the command line, or inside a continuous integration pipeline. To get started with an account, begin with Dashboard — Overview; for a one-off test without an account, use the site playground.

One product, several analyses

The dashboard brings together several independent analyses. Use the ones that match your site and goal; Agent Commerce applies only to online stores.

AnalysisQuestion it answers
Agent ReadinessCan an agent find, read and use your site, whatever your industry?
Agent ExperienceWhat happens when the results are reframed as seven understandable questions?
Agent PerformanceDoes your site stay readable, stable and fast enough during agent navigation?
Agent Commerce (stores)Does the store expose a catalog and buying journey an agent can use?
Brand DiscoveryDo your official resources appear when an agent searches for your brand online?
The simulation stops short of payment
No payment is ever triggered. The Agent Commerce simulation may create a cart and open an order to observe how far an agent would get, then stops there and abandons the session. No card data is handled at any point.

The foundation: general Agent Readiness

On top of those three checks, an audit can run an independent scan of 27 signals — whether an AI agent can discover, read and use the site, whether it sells anything or not. Fourteen signals determine the tier; thirteen diagnostics add context. They are grouped into discoverability, technical connection, content accessibility and bot access control.

This is the general_scan option, detailed in the API reference. It produces its own result, general_readiness, with its own maturity tier — not ready, almost ready, agent ready, fully optimized — read separately, never blended into overall_score.

Reading results

ScoreStatusWhat it means
85 – 100passAn agent can use this site without meaningful obstacles.
60 – 84warnUsable, but with blind spots that will degrade the agent's experience.
0 – 59failA structural obstacle stops the agent from doing its job. The recommendation says which.

Compatibility

For stores, Agent Commerce builds on UCP (Universal Commerce Protocol), an open, platform-agnostic standard. This analysis is one part of the product, not the definition of agent-readiness for a general website.

In practice: general Agent Readiness works on any domain. If you run a store exposing a valid UCP profile, you can also run Agent Commerce, whatever technology powers it.

Where to start

  • First call in under five minutes: the Quickstart page.
  • Wiring it into a pipeline with a blocking threshold: the CLI page, then CI integration.
  • Handling errors client-side: the Error contract page.
  • A machine-readable contract: the OpenAPI spec at /openapi.json, listed in the API catalogue at /.well-known/api-catalog, alongside the auth flow described at /auth.md.