Niche.io analyzes your YouTube channel and matches you with brands that are the right fit — then writes your pitch email for you.

Built at HOTH XIII with Next.js, Anthropic API, and the YouTube Data API.

1. You enter your YouTube channel — the app pulls your real subscriber count, average views, and recent video titles
2. Pick your creator niche — lifestyle, fitness, beauty, food, tech and more
3. Get 5 brand recommendations ranked by compatibility:
   • 2 Reach brands — iconic names that are a stretch but worth shooting for
   • 2 Target brands — best-fit brands with high acceptance odds
   • 1 Related brand — a brand from a similar niche to broaden your options
4. Claude writes your pitch — each recommendation includes a personalized creator bio and cold-pitch email ready to send

git clone https://github.com/KayKimKwon/Content-Creator-Portfolio
cd Content-Creator-Portfolio/media-kit
npm install

Create a .env.local file in the media-kit/ folder:

ANTHROPIC_API_KEY=sk-ant-...
YOUTUBE_API_KEY=AIza...

• ANTHROPIC_API_KEY — required for AI-generated bios and pitch emails. Get one at console.anthropic.com
• YOUTUBE_API_KEY — optional but recommended. Without it, the app estimates your stats from Instagram/TikTok followers instead. Get one from Google Cloud Console

npm run dev

Open http://localhost:3001.

Each brand is scored against your profile across three factors:

Brand fame (0–1 scale) is then blended in at 40% weight for the final score. Brands you've worked with before get a -10 penalty; brands you listed as targets get a +8 boost.

Acceptance probability is bucketed as High (score ≥ 80), Medium (60–79), or Low (< 60).

Pricing estimates are calculated as avg views × brand CPM / 1000, with a scaling curve for large creators.

media-kit/
├── src/
│   ├── app/
│   │   ├── page.tsx          # Landing page
│   │   ├── input/page.tsx    # Creator input form
│   │   ├── output/page.tsx   # Brand recommendations
│   │   ├── stats/page.tsx    # Creator stats view
│   │   └── api/
│   │       ├── match/        # Core matchmaking + AI pitch generation
│   │       ├── meta/         # Returns available niches
│   │       ├── analyze/      # Creator analysis
│   │       └── generate/     # Standalone pitch generation
│   └── lib/
│       ├── brands.ts         # Loads brand data from per-niche JSON files
│       └── youtube.ts        # YouTube Data API integration
└── data/
    ├── meta.json             # Niche list, creator tier config
    ├── similar-niches.json   # Related niches map
    └── niches/               # Brand data per niche
        ├── gaming.json
        ├── fitness.json
        ├── beauty.json
        └── ...

Each niche has its own JSON file under data/niches/. To add a brand:

[
  {
    "name": "BrandName",
    "brandTier": "iconic",
    "fameScore": 0.9,
    "idealCreatorTier": ["micro", "mid"],
    "minSubscribers": 10000,
    "maxSubscribers": 500000,
    "avgCPM": 25,
    "collaborationStyle": "dedicated video",
    "brandPersonality": "innovative, youthful"
  }
]

To add a new niche, create data/niches/<niche>.json and add the niche name to the niches array in data/meta.json.

Built at HOTH XIII (UCLA Hackathon) — March 2026.