v1.6.0

OpenMOSS/MOSS-Audiov1.6.0Apr 20, 2026by shrimbly

AI Summary

This release enhances the Seedance 2 image-to-video nodes by supporting richer media inputs like reference images, videos, and audio. It also resolves a critical bug preventing reference-only workflows from executing correctly.

Key Highlights

Enhanced Seedance 2 I2V nodes with support for reference images, videos, and audio inputs.
Fixed reference-only runs that were previously rejected by the backend.
Documented mutual exclusivity rules for First/Last Frame vs Reference inputs.

New Features

Seedance 2 I2V: richer media inputs (Last Frame, Reference Images, Reference Videos, Reference Audio)

Full Release Notes

### Added

- **Seedance 2 I2V: richer media inputs** — The ByteDance Seedance 2.0 and 2.0 Fast image-to-video nodes now expose Last Frame, Reference Images (up to 9), Reference Videos (up to 3), and Reference Audio (up to 3) handles alongside First Frame. Handle descriptions document the First/Last Frame vs Reference Images mutual-exclusivity rule.

### Fixed

- **Seedance 2 I2V: reference-only runs no longer rejected** — When connecting images only to Reference Images, the request no longer duplicates them into `first_frame_url`, which Kie was rejecting as a mutually-exclusive combination.