MoonPay Payroll

CSV Schema

Required headers:

• name
• chain_address
• token_name
• chain_name
• account
• frequency

Validation Rules

• Header set must match required payroll schema.
• chain_address must be valid EVM format for current implementation.
• Unknown token/chain mapping is flagged during planning.
• Invalid rows are marked and skipped from executable routes.

Planning Model

1. Parse CSV
2. Validate rows
3. Resolve treasury position (account/token/chain/balance)
4. Generate per-row settlement route:

• direct
• swap
• bridge
• bridge+swap

Planner is reusable and available in shared MoonPay/payroll modules.

Execution Model

• Batch rows are persisted with row statuses:
  • pending, validated, routed, executing, completed, failed, skipped, awaiting_user_action
• Execution is idempotent per batch/row.
• Row failures do not abort unrelated rows.
• Failed rows can be retried.

Scheduling

Scheduler state is persisted in SQLite:

• payroll_batches
• payroll_batch_rows
• payroll_schedules
• payroll_run_history
• payroll_accounts

Run due schedules:

pnpm payroll:run-due

Notes

• Payroll is treasury-funded and must not bypass treasury controls.
• Buy-crypto flows are user-assisted and not the primary autonomous payroll path.
• Autonomous paths prefer swap/bridge/transfer routes.