Skip to main content
Each prompt gives your coding agent the endpoints, enum values, and constraints for one Lumx task, so it builds against the API instead of guessing. Paste a prompt into Claude Code, Cursor, Codex, or any agent that can read a URL, and it reads these docs and the OpenAPI spec before writing code. Every prompt follows the same shape: the sources to read first, numbered steps with literal endpoints, constraints, and the deliverables to expect. No prompt creates an API key, so create your Sandbox key before you start. See Authentication.
Connect the docs MCP server before you paste a prompt. The agent can then look up any field it is unsure of instead of relying on memory.

Setup and testing

Get to a working Sandbox integration, then harden it for production.

End-to-end integrations

Build one Lumx capability from the first call to settlement.

Build a product

Combine several capabilities into a complete payment product.

Migrate from an existing setup

Move a payment flow you already run onto Lumx.

Ship with AI

Builder guides, the docs MCP, and the API.

MCP server

Connect AI tools to these docs.