AI Video Editor is a detailed BeforeCode reference pack for a browser-based AI video creation and editing platform. It is inspired by advanced product documentation patterns: product vision, PRD, TRD, workflow orchestration, AI agents, rendering, storage, API contracts, QA, and handoff.
The product helps creators and teams generate, edit, review, and export short-form videos from scripts, uploads, templates, captions, and AI-generated scene suggestions.
User creates workspace project
→ uploads or writes source content
→ AI suggests scenes and captions
→ user edits timeline
→ render job is queued
→ preview is reviewed
→ final export is generated
| File | Purpose |
|---|---|
00-project-context.md |
Source-of-truth product context |
01-open-questions.md |
Product, technical, and rights decisions |
03-prd.md |
Product requirements and user workflows |
04-trd.md |
System architecture and module boundaries |
05-system-architecture.md |
C4-style architecture, services, and lifecycle |
06-agent-workflows.md |
AI planning, caption, scene, and review workflows |
07-rendering-pipeline.md |
Queue, FFmpeg/export, presets, failure states |
08-qa-regression-plan.md |
Regression suite for editor/rendering/AI |
09-ai-coding-handoff.md |
Safe implementation instructions for AI agents |
- How to document a complex AI product before implementation.
- How to separate editor UX, AI workflow, media storage, and rendering.
- How to define human review before export.
- How to write QA plans for timeline, captions, rendering, and failure states.
- How to prevent AI agents from inventing rights, rendering, or storage behavior.
npx beforecode start --from examples/ai-video-editor/00-project-context.md