Sync Labs · Rate Limits

Sync Labs Rate Limits

Sync Labs enforces two distinct kinds of limits. A request-rate limit applies per API key (60 rpm for POST /v2/generate, 600 rpm for everything else) and returns 429. A separate concurrency limit caps simultaneous PENDING/PROCESSING generations and is gated by subscription tier (1 / 3 / 6 / 15 / Custom). Both surface as 429 but require different remediation - rate-limit waits clear within a minute while concurrency requires a slot to free.

Sync Labs Rate Limits is the machine-readable rate-limit profile for Sync Labs on the APIs.io network, conforming to the API Commons Rate Limits specification.

It captures 7 rate-limit definitions, measuring requests_per_minute and concurrent_requests.

The profile also includes 4 backoff/retry policies defined and response codes documented for throttled.

Tagged areas include Artificial Intelligence, Lip Sync, Video, and Rate Limiting.

7 Limits Throttle: 429
Artificial IntelligenceLip SyncVideoRate Limiting

Limits

Generation Endpoint Rate Limit key
requests_per_minute · minute
60
Applies to POST /v2/generate. Resets each minute.
Other Endpoints Rate Limit key
requests_per_minute · minute
600
Applies to all non-generation endpoints (list, get, asset management, etc.).
Concurrent Generations - Hobbyist account
concurrent_requests
1
Maximum simultaneous PENDING/PROCESSING jobs on the Hobbyist plan.
Concurrent Generations - Creator account
concurrent_requests
3
Maximum simultaneous PENDING/PROCESSING jobs on the Creator plan.
Concurrent Generations - Growth account
concurrent_requests
6
Maximum simultaneous PENDING/PROCESSING jobs on the Growth plan.
Concurrent Generations - Scale account
concurrent_requests
15
Maximum simultaneous PENDING/PROCESSING jobs on the Scale plan.
Concurrent Generations - Enterprise account
concurrent_requests
custom (negotiated)
Enterprise concurrency is negotiated per-contract.

Policies

Exponential Backoff
The docs recommend exponential backoff (1s, 2s, 4s, 8s, 16s using 2^attempt). Do not retry 429 responses immediately in tight loops.
Rate vs Concurrency Distinction
A 429 from a rate-limit hit clears within a minute; a 429 from concurrency requires existing jobs to complete. Clients should differentiate the cause and back off accordingly.
Concurrency Monitoring
Use the List Generations endpoint to track active jobs, implement client-side queuing, and use webhooks for completion notifications instead of polling.
Batch API
Scale and Enterprise tiers gain access to the Batch API which automatically manages concurrency across larger workloads (up to 500 videos per batch).

Sources