Build with an AI agent
A ready prompt for Cursor, Claude Code or any coding agent: it reads llms.txt and the OpenAPI document and integrates Norelio by the rules that matter — the key on the server, payments created server-side, verified webhook signatures, every event processed once.
Your coding agent — Cursor, Claude Code or any agent that reads documentation — can do the integration. Paste the prompt below into the agent in your project. It sends the agent to llms.txt, the index of this documentation, and to the OpenAPI document, and names the rules it must not get wrong: the key stays on the server, payments are created server-side, webhook signatures are verified over the raw body, every event is processed once.
You are connecting my application to Norelio. Norelio is the merchant of record: it takes the payment and pays me out.
Read https://docs.norelio.eu/llms.txt and the pages it links to. The OpenAPI document is https://docs.norelio.eu/openapi.json.
Ask me for the Norelio API key and the productId of each product. Keep the key on the server, in the environment variable NORELIO_API_KEY. Never put it in client code or logs.
When a user presses "Pay", create the payment from my server: POST https://app.norelio.eu/api/v1/checkouts with the header Authorization: Bearer <key> and the fields productId and clientReference (my id for that user). For a top-up product also send packageId or amount. Redirect the user's browser to the url in the response; it works once, so do not log it.
If the answer is 403 direct_checkout_not_enabled, stop and tell me: creating payments from my server has to be enabled for my account first.
Add a webhook endpoint and tell me its address: I save it in the Norelio cabinet. Verify the signature over the raw request body, deduplicate on eventId, and on order.paid credit the account named in clientReference (for a top-up, add the credits field). Handle refunds. For subscriptions, call the access check whenever the user comes back.
Tell me what you changed and how to test it with the test webhook from the Norelio cabinet.The prompt asks the agent to request the API key (the partner issues it in the cabinet under API and webhooks → API key) and the id of each product (the cabinet catalog), to name the webhook endpoint it added — the partner saves that address under API and webhooks → Webhooks — and to explain how to test with the test notification from the cabinet. Review the change as you would a colleague's.
What the agent reads
- llms.txt
- The index: one line per section, with links. The prompt sends the agent here first.
- OpenAPI 3.1
- The endpoints and every webhook payload as a machine-readable contract. The prompt names it too.
- llms-full.txt
- This whole documentation in one Markdown file, every code example included, for an agent that prefers a single file.
To check the agent's work, follow the Quickstart: it shows the same integration step by step, in the language of your project.
