first-last-frame

Seedance 2.0 ComfyUI: custom nodes and workflows for ByteDance video generation via MuAPI — text-to-video, image-to-video, Omni Reference, consistent characters, and video extend.

Local ComfyUI video generator for first/last-frame animation and synchronized audio with Wan2.2 and MiniMax H3, verified on RTX 5090.

一键提取视频首帧、尾帧与任意指定帧 · 纯浏览器本地处理 · 零依赖 · 不上传 免注册 无水印

Frame chaining for AI video: seamless multi-clip continuity for image-to-video (ComfyUI adapter, stitch & speed-ramp included)

A field manual for image-to-video production: frame extraction, first/last-frame pairing, prompt structure, ffmpeg-only repair and quantitative motion checks.

AI 广告视频实战案例：一天做出一支 36 秒口红广告，每一步、每一版修改都记录在案 | AI video ad case study: GPT Image 2.5 + MiniMax H3 + ffmpeg — storyboard, keyframes, first-last-frame shots, editing, beat sync

Turn a video link or local file into a shot-by-shot breakdown (shot list, first/last frames) and per-shot AI video generation jobs, with an evidence gate and a zero-cost dry-run. Inspired by REA (Reverse Engineer Anything); complementary to rea-video.

基于 MiniMax H3 API 的本地导演工作台：文生视频 / 首尾帧生视频 / 视频生视频 + 音画同生，云端算力渲染、与本地硬件无关，成片自动归档本机。 | Local director workbench on the MiniMax H3 API: text / first-last-frame / video-to-video with audio, cloud rendering independent of local hardware, results archived on your machine.

Official-template companion for Seedance 2.0 ComfyUI T2V, R2V and FLF2V workflows, prompts, setup, JSON checks and troubleshooting.