CapCut's new Hub merges AI video generation and editing into one workflow
量子位 · wechat · 2026-09-20
At the Douyin Creator Conference, CapCut (Jianying) launched major updates that bridge AI generation and editing.
On PC: the new Hub generates images (Seedream 5.0 Pro), storyboards, assets, and multi-shot videos (Seedance 2.5) in one canvas, then hands off to CapCut's multitrack editor. A CapCut Assistant Agent edits timelines via text commands with 20+ official Skills (talking-head cuts, narration, subtitle fixes, batch output), plus AI clip extension, local edits, upscaling, frame interpolation and color tools. Assets from Jimeng and other ByteDance tools import in one click.
On mobile: snap-to-video generates Vlogs from raw photos plus a prompt, with the same assistant and skills available. The reviewer's takeaway: CapCut is consolidating the whole creation pipeline rather than adding isolated AI features.
More from Multimodal
- Was Midjourney quietly upgraded? Striking stereoscopic portraits on v8.2 — michaelrabone · 2026-09-20
- Claude beats GPT image 2.5 at on-brand design 9/10 times with a brand guide — sahilypatel · 2026-09-20
- Filmic Plus Turns Your Blender Scene Into a Real Fujifilm Camera — ssh4net · 2026-09-20
- YuE2 in ComfyUI: open-source music generation with piano roll beats Suno, claims demo — Maleficent-Tell-2718 · 2026-09-20
- Users struggle to make Minimax Ref2Va preserve reference animation like Beeble — galo64 · 2026-09-20
- HKUST-led team open-sources YuE2 music model that beats Suno v6 on WildSongBench — 量子位 · 2026-09-20