Agent-readable docs
These resources describe the same target contract as the guides, in formats that AI coding assistants and agents can read without rendering the site.
| Resource | Format | Contents |
|---|---|---|
/llms.txt |
Text (llms.txt convention) | Summary, key rules and links to everything below |
/openapi-target.json |
OpenAPI 3.1 JSON | Target contract: paths, schemas, examples, errors and webhooks |
| Licensing requests | Markdown | Request, response, idempotency and errors |
| License verification | Markdown | Fail-closed decision rule and fields |
| Licensing lifecycle | Markdown | Events and webhook signing |
| MCP & agent access control | Markdown | Identity vs license responsibilities |
Target vs production documents
Section titled “Target vs production documents”/openapi-target.json is the target contract that implementations conform to. It is deliberately
named differently from any future production /openapi.json, which will describe only deployed
behavior. Do not generate production clients from the target document without checking the changelog.
Guidance for agents
Section titled “Guidance for agents”When generating integration code from these documents:
- Keep the API key, webhook secret and onboarding links in server-side configuration only.
- Treat only
valid === trueas permitting access; everything else denies. - Use the
licenceKeyfromlicence.issuedexactly as delivered. - Keep end-user authentication in the Seller’s own identity provider.