Start here
Environment, context files, and IDE setup.
Prompt library
Copy-ready prompts for collect, payout, debug, and review.
AI & MCP
llms.txt, documentation MCP, and agent checklist.
Quick Start
Human-readable first collection walkthrough.
Rules every session should follow
- Server-side only —
DOLLR_CLIENT_SECRETnever in browsers, mobile apps, or public repos. - Attach context — llms.txt plus the relevant stack guide or Quick Start.
- One flow per task — collect or payout or checkout; not the entire API at once.
- Idempotency — UUID v4
reference_idbeforePOST /v1/executions/*; poll status before retrying. - Two status models — execution status vs invoice/order status; MoMo may stay
PROCESSINGfor minutes. - Markets — use Payments by market; predict MoMo operator from phone when possible.
When to use prompts vs guides
| Use | When |
|---|---|
| Prompt library | You want the model to generate or refactor integration code |
| Integrate by stack | You want step-by-step reference while coding |
| API Reference | You need exact request/response fields |
