Rate limits
Rate limits are enforced per connected system, so a slow platform never throttles a fast one.
Because unified calls fan out to real platform APIs, rate limits are enforced per connected
system — a slow platform never throttles your calls to a fast one. When a limit is hit, Flow
queues briefly and then returns 429 Too Many Requests; back off and retry with exponential
delay.
| System | Auth | Rate limit |
|---|---|---|
| Shopify | OAuth 2.0 | 1 req / s |
| Lightspeed X-Series | OAuth 2.0 | 70 req / 6s |
| Lightspeed R-Series | OAuth 2.0 | 1 req / 10s |
| Square | OAuth 2.0 | 5 req / s |
| Clover (Asia Pacific) | API token | 5 req / 6s |
| Toast | API token | 5 req / s |
| Clover (US) | API token | 5 req / 6s |
| Lightspeed K-Series | OAuth 2.0 | 100 req / 60s |
| QuickBooks Online | OAuth 2.0 | 50 req / 6s |
| Xero | OAuth 2.0 | 1 req / s |
| Cin7 Core | API key | 2 req / 3s |
Per-entity limits can be stricter than the system-wide default — each entity's Supported systems table in the API reference lists the exact limit that applies to its reads and writes.