Skip to main content
The sandbox gives you three endpoints to manage test data for your merchant:
  • POST /v1/merchant/sandbox/reset — wipe and optionally reseed
  • POST /v1/merchant/sandbox/seed-beneficiaries — generate sample beneficiaries across multiple countries and methods
  • POST /v1/merchant/sandbox/seed-payouts — generate sample payouts in mixed states for UI and webhook testing
All three are sandbox-only — calling them against production returns 403 sandbox_only. They’re scoped to your merchant — you can’t affect another merchant’s data.

​
Reset

Four reset scopes:
?scope=What it does
balances (default)Wipe merchant_accounts + sandbox transactions, re-seed fresh balances. Beneficiaries, instruments, and payouts stay.
allWipe payouts, instruments, beneficiaries, and sandbox transactions. Re-seed currency balances only.
delete-allWipe everything. No reseed. You land at zero state.
full-resetWipe everything. Re-seed balances + sample beneficiaries + sample payouts. Equivalent to a fresh merchant with demo data.
# Clean balances, keep your beneficiaries and payouts
curl "https://api.antonpayments.dev/v1/merchant/sandbox/reset?scope=balances" \
  -X POST \
  -H "Authorization: Bearer $ANTON_API_KEY"

# Nuclear option — fresh merchant state with demo data
curl "https://api.antonpayments.dev/v1/merchant/sandbox/reset?scope=full-reset" \
  -X POST \
  -H "Authorization: Bearer $ANTON_API_KEY"
Rate-limited to 10 resets per merchant per hour — enough for active development, tight enough to catch a runaway test suite. Response: 
{
  "status": "reset",
  "scope": "balances",
  "merchant_id": "mer_01HX..."
}
delete-all responds with "status": "deleted".

​
Seed sample beneficiaries

Creates a curated set of realistic beneficiaries across multiple countries and payment methods — useful for populating your UI or testing country-specific edge cases: 
curl https://api.antonpayments.dev/v1/merchant/sandbox/seed-beneficiaries \
  -X POST \
  -H "Authorization: Bearer $ANTON_API_KEY"
Response: 
{
  "status": "seeded",
  "created_count": 12
}
The beneficiaries created include a mix of individual and business types, spanning several corridors — enough variety to exercise most payout flows without you having to construct PII by hand.

​
Seed sample payouts

Creates payouts in a mix of states — completed, failed, cancelled, manual_review, and in-flight — so you can validate your UI and webhook handling against every outcome: 
curl https://api.antonpayments.dev/v1/merchant/sandbox/seed-payouts \
  -X POST \
  -H "Authorization: Bearer $ANTON_API_KEY"

{
  "status": "seeded",
  "created_count": 30
}
Requires seeded beneficiaries first. Running seed-payouts with no beneficiaries returns 422 no_beneficiaries. Call seed-beneficiaries first, or use ?scope=full-reset on the reset endpoint which seeds both.

​
Using these in CI

A recommended CI pattern:
1

Before each test run, reset

POST /v1/merchant/sandbox/reset?scope=delete-all to guarantee a clean slate.
2

Seed any baseline data your tests assume

If your tests need a starting population (e.g., dashboard snapshot tests), call seed-beneficiaries and seed-payouts.
3

Run your tests

Exercise your real integration against the seeded state.
4

Between test suites, reset scoped to balances

?scope=balances is faster than a full reset — keeps your seeded data, resets funding state.
Rate limits apply across the hour: 10 resets per merchant per hour. In a heavy-test CI setup, either share a merchant across workers carefully, or provision multiple sandbox merchants.

​
Triggering specific scenarios (future)

Sandbox seeders produce a curated distribution — payouts in mixed states, beneficiaries across corridors. For deterministic triggering of specific engine verdicts, screening hits, or rail failures, that’s forthcoming functionality — when it lands it’ll use the ?simulate= query parameter on POST /v1/payouts (sandbox-only). Watch the changelog for when it ships.

​
What’s different from production

See Sandbox Overview for the full list. Key points:
  • Rail submission is mocked — no real banks are contacted
  • Timing is compressed — sandbox payouts transition to completed in seconds
  • Real OFAC screening runs, but you can trigger deterministic outcomes using specific test beneficiaries
  • All authentication, authorization, webhook signing, and state-machine behavior is identical to production

​
Next steps

Sandbox overview

What sandbox is, and how it differs from production.

Going live

Checklist for promoting from sandbox to production.