待編排5 生成中8 已完成12
33 個模型 · 同步中 2026 / 05 / 20 · 09:51
JC
姜承彥
技術製片

把這5條指令,編排成5條工作流。

來自 春季品牌 TVC · 創意台派發 · 18 分鐘前
NT$847
總成本
2h 15m
總時長
17 節點
跨 5 條工作流
01 5 條工作流 drag to reorder
01 SHOT 待編排
SC-01 開場鏡頭重出
NT$187 45m 5 節點
02 PACE 生成中
情緒段每鏡 +0.4~0.6s
NT$75 22m 3 節點
03 COPY 已完成
Slogan L-04 前置至 04–05s
NT$0 5m 2 節點
04 GRADE 待編排
v2.0 暖調 LUT · 6200–6400K
NT$250 38m 4 節點
05 CUT 待編排
對照剪輯(近第一支樣片)
NT$335 25m 6 節點
本日總計
5 條工作流 · 17 節點 NT$847
預估完成 11:32 2h15m
02 Production POC Milestone 02 · OpenAI image generation

Production Station POC

Upload the breakdown script and style reference image, then trigger real per-cut image generation through the Production POC Worker.

Worker image flow
Inputs 2 required files
Breakdown Script
Send extracted script text to the Worker, which remains the single parser for cut generation.
Pending
No file selected
Style Reference Image
Accept the visual style reference that later image generation will follow.
Pending
No file selected
Automation trigger
Extracts script text in-browser, sends it with the reference image to the Worker, and requests per-cut image generation.
Upload both required files to enable automation.
Workflow 5 milestone steps
  1. 01
    Upload breakdown script
    Required before the script can be parsed into cuts.
    Pending
  2. 02
    Upload style reference image
    Required before the automation sequence can start.
    Pending
  3. 03
    Parse cuts from script
    The browser extracts raw text, then the Worker parses the script into normalized cuts.
    Locked
  4. 04
    Generate image per cut
    The Worker sends each cut plus the reference image to OpenAI and returns generated image assets.
    Locked
  5. 05
    Generate final video
    Still deferred. This milestone does not call Seedance or create final video outputs.
    Locked
Downloads Session-only outputs
Generated Images
Downloads the generated cut images returned by the Worker. Files stay in memory only for this page session.
Not ready
Generated Video
Not part of this milestone. Video generation remains deferred until the next integration stage.
Deferred
No generated images in this session.
Generated cut images will appear here after the Worker completes image generation.
Milestone Scope What is intentionally missing
This milestone uses browser-side text extraction and a Cloudflare Worker for real OpenAI image generation. It still excludes permanent storage, auth, database work, R2, and Seedance video generation.
希拉雅 Siraya · 製片台 · 模塊 03 / 06 · AI 工作流編排
v2.6.0 built for people who actually make ads.