Apr 18, 2026
v1.2.7.19
Manual video split editor, face mosaic, drawing tools, text note titles, board performance

Related tutorials
8:17Level Up Your AI Producion | Flashboards & Seedance 2.0
Watch on YouTube
12:25Seedance 2.0 Omni Complete Workflow | Total Control
Watch on YouTube
13:06Complete Seedance 2.0 Video Reference Tutorial | Motion, Camera & Style
Watch on YouTube
10:56I Turned Myself into a 3D Movie Character! | Seedance 2 AI Tutorial
Watch on YouTube
Models & workflows
New (2)
Seedance 2 Official — official Seedance 2 video generation with optional image, video, and audio reference inputs
Seedance 2 Official Fast — faster official variant at lower cost with the same multi-reference support
Updated (3)
Seedance 2 Omni — now supports combined image + video + audio references in a single generation
Kling O3 — fixed FPS preflight detection
Sora 2 Basic & Pro (T2V, I2V) — flagged as Under Review due to upstream issues
What's new
- Multi-select ZIP download — select multiple board items and download them as a single ZIP file
- Text note titles — add titles to text notes with customizable border colors, style presets, and a new font (Zalando Sans Expanded)Open
- Board media info popover — hover over items to see file details like dimensions, size, and formatOpen
- Drawing tool improvements — expanded toolbar with opacity controls and keyboard shortcuts
Improvements
- Board performance — low-zoom and dense-view simplification for smoother interaction with large boards, per-note toolbars suppressed during multi-selection
- Board grid now renders as an infinite viewport overlay with smooth fade based on zoom level
- Media browser opens faster and handles expired audio and video previews gracefully
- Input images on Explore model pages are now clickable
Fixes
- Fixed curated collection switching in media browser
- Fixed multiselect delete in overlay mode and batch delete keyboard shortcut
- Fixed connection overlay drag scope sync
- Fixed board generation results being lost after removing the source item
- Fixed duplicated, copied, or split media inheriting AI input assignments from the original

