SpeakBand Studio MVP V9
AI Auto Edit workflow for short-form IELTS promos.
Upload source clips, analyze metadata, describe what you want, choose a target duration before AI Auto Edit, preserve that AI-planned timeline through FFmpeg command metadata, and simulate a render worker lifecycle.
V4 STEP 1
Upload Video
Add source clips for AI analysis, future auto editing, and eventual MP4 generation. Files upload to Cloudflare R2 while local object URLs stay available for browser previews.
Uploaded clips
No clips yet
No clips uploaded yet. Add source footage to prepare for V4/V5 editing.
V4 STEP 2
AI Material Understanding
Uses uploaded clip metadata to catalog source material before the brief-writing step. If a Video Idea / Instruction or Scene Outline already exists, it can add context, but this step does not require either field and does not optimize for a target MP4 duration.
V4 STEP 3
Create Brief
Describe what you want the edit to achieve. This replaces the old separate description input and marketing-generator fields.
Optional but recommended for screen recordings or product demos. Tell AI what happens at each timestamp so it can choose better moments.
V4 STEP 4
AI Auto Edit
Uses uploaded clip metadata, the Video Idea / Instruction, the selected Target Duration, and the AI Analysis result to generate an execution-ready timeline JSON with numeric trim and placement seconds optimized for the selected final MP4 length.
V9 STEP 5
Render Job System
Convert the AI execution timeline into a render plan without creative retiming, choose Simulation or Real Render direction, then keep Cloudflare Pages on the edge-compatible simulation and R2 artifact workflow.
Render Plan preserves the AI timeline selected in Step 4; target duration is first used by AI Auto Edit before timeline generation.
Generate the V4 Step 4 execution timeline before creating a render plan and render job scaffold.