Chameleon uses AI to seamlessly integrate sponsor products directly into creator videos — making ads feel like a natural part of the content rather than an interruption. Brands submit their product, AI identifies the optimal placement window in the video, and Runway ML generates a photorealistic edited version with the product blended in.

Built at CUHackit 2026

Traditional sponsorships are jarring — creators stop mid-video to hold up a product. Viewers skip it. Brands get low engagement. Chameleon makes sponsorships invisible: the product appears naturally in the scene, at the right moment, without breaking the flow of the content.

1. Creator uploads a video — stored directly in S3
2. TwelveLabs AI analyzes the video — extracting topics, hashtags, sentiment, and identifying the optimal ad placement timecodes
3. Brands browse and make offers — attaching their product image and budget
4. Creator accepts — triggering the automated editing pipeline
5. Runway ML generates the sponsored version — the product is integrated into the scene at the AI-identified placement window, preserving lighting, motion, and context
6. Both parties view the result — the final edited video appears on the Sponsored Videos page

Creator uploads video
        │
        ▼
    Amazon S3
        │
        ▼
  EventBridge rule
        │
        ├──► Lambda: chameleon-video-analyzer
        │         └─ TwelveLabs → topics, hashtags, sentiment → DynamoDB
        │
        └──► Lambda: chameleon-ad-analyzer
                  └─ Detects optimal ad placement timecodes → DynamoDB

Brand makes offer (with product image) → FastAPI → DynamoDB
Creator accepts → FastAPI → EventBridge (OfferAccepted)
        │
        ▼
Lambda: chameleon-video-edit
        └─ Downloads video + product image from S3
        └─ Runway ML seamlessly integrates product at the AI-identified timecode
        └─ Uploads edited video to S3
        └─ Updates DynamoDB (editStatus=complete)

Chameleon/
├── frontend/               # React app (Vite + Tailwind)
│   └── src/
│       ├── pages/
│       │   ├── Landing.tsx
│       │   ├── CreatorDashboard.tsx
│       │   ├── UploadVideo.tsx
│       │   ├── CompanyBrowse.tsx
│       │   ├── VideoDetail.tsx
│       │   └── SponsoredVideos.tsx
│       ├── components/
│       │   ├── Navbar.tsx
│       │   ├── PageShell.tsx
│       │   ├── VideoCard.tsx
│       │   └── ProtectedRoute.tsx
│       └── lib/
│           └── api.ts
├── backend/
│   ├── main.py             # FastAPI app (Elastic Beanstalk)
│   ├── handler.py          # Lambda 1: TwelveLabs video analysis
│   ├── ad_analyzer.py      # Lambda 2: Ad placement timecode detection
│   └── video_edit_lambda.py # Lambda 3: Runway ML video editing
└── infra/
    └── integration.md      # AWS infrastructure guide

• AI placement detection — TwelveLabs identifies the exact moment in a video where a product can be naturally inserted
• Runway ML integration — photorealistic product placement that matches the scene's lighting, motion, and context
• Product image uploads — brands attach their product photo when making an offer
• End-to-end automated pipeline — from offer acceptance to finished edited video with no manual steps
• Role-based access — separate creator and brand experiences
• Direct S3 uploads — videos never pass through the backend server
• Secure media delivery — all video playback via time-limited S3 presigned URLs