ModelHub

High-Availability Seedance Video Generation API Relay

ModelHub is a unified, OpenAI-compatible gateway for Seedance 2.0, 2.0 Fast, 2.0 Mini and Seedance 2.5. Point your existing client at one Base URL and submit text-to-video, image-to-video, first/last-frame and omni-reference jobs. Tasks are routed across multiple upstream providers to the lowest-cost node that satisfies the requested ratio, resolution and duration, with automatic failover when a provider fails.

Featured Seedance video models

  • Seedance 2.5 (seedance-2.5): Next-gen flagship model. Supports 4-30s video generation with omni-reference over up to 50 assets. 480p / 720p / 1080p, 4-30s, Omni-Reference (50 assets).
  • Seedance 2.0 (seedance-2.0): Director-grade realistic physical generation with refined camera motion, plus first/last frame transitions. 720p / 1080p, 4-15s, First/Last Frames.
  • Seedance 2.0 Fast (seedance-2.0-fast): A balance of speed and quality: rapid iteration with reliable cinematic composition and movement. 720p, 4-15s, Fast Turnaround.
  • Seedance 2.0 Mini (seedance-2.0-mini): Cost-efficient lightweight model with 2x generation speed, the pick for large-scale batch production. 720p, 4-15s, 2x Speed.

Core advantages

  • Least-cost smart routing: Compares upstream providers on price and availability and dispatches each task to the cheapest viable node, saving 30%-50%.
  • Omni-reference mode (omni_reference): Drive a generation with up to 30 images, 10 videos and 10 audio clips, 50 assets in total.
  • First & last frames (first_last_frames): Lock a shot transition between a start and an end frame for continuous character and scene motion.
  • OpenAI-compatible: Standard Bearer-token auth and JSON payloads, no client rewrite required.

Get started

  • API docs: Base URL https://modelhub.cc/v1; POST /v1/videos/generations to submit, GET /v1/videos/tasks/{id} to poll.
  • Recommended skills: Ten open-source Seedance video skills, rendering unified on ModelHub.