What Are Seedance 2.5 API Rate Limits?

Overview

Seedance 2.5 API rate limits should be treated as account- and route-dependent unless your current WaveSpeedAI dashboard or documentation states a specific number. This batch’s facts do not support publishing a universal requests-per-minute or concurrency limit for every user.

Design as if limits will matter. Put generation requests into a queue, apply per-user and per-workspace ceilings, use backoff on throttling, and keep idempotency keys or duplicate protection so retry storms do not create double submissions.

Rate limits are not only about HTTP responses. Video generation can bottleneck on queue depth, duration, references, moderation, storage, or account spend controls. A reliable product should track pending, running, completed, failed, and timed-out jobs separately.

WaveSpeedAI gives a standard API route for submitting predictions, but your product should still avoid hard-coding assumptions. Keep rate-limit handling configurable, log provider responses, and expose friendly user messages when a job is delayed or rejected by current account limits. Enterprise customers may also need separate workspace-level throttles, especially when several teams share one billing account. Treat throttling as normal capacity planning, not an outage by default.

Source: ByteDance Seed official Seedance 2.5 page; WaveSpeedAI Seedance 2.5 API collection. Source check: 2026-08-29.