fluencerfinder

Multi-platform creator / influencer discovery and classification engine. Search by niche, topic, hobby, or location and get back creator handles with LLM-generated classifications across Instagram, TikTok, X, and YouTube.

Inspired by RightFluencer's multi-platform pipeline approach, rebuilt with FastAPI, SQLModel, database-first search, and structured LLM categorization.

Repository layout

.
├── creator-discovery/     # The application (FastAPI backend + Vite/React frontend)
├── worker.js              # Cloudflare Worker: serves the SPA + same-origin API proxy
├── wrangler.toml          # Cloudflare Worker / Pages config
├── render.yaml            # Render blueprint for the FastAPI backend
└── Dockerfile             # Container image for the backend

The application lives in creator-discovery/ — see its README for full setup, API reference, and deployment docs.

Quick start

cd creator-discovery

# Backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

# Seed the bundled sample dataset
python scripts/seed_data.py

# Run the API
uvicorn app.main:app --reload --port 8000

In another terminal:

cd creator-discovery/frontend
npm install
npm run dev

Open http://localhost:5173 and search for "Los Angeles fitness creators".

No API keys are required for basic discovery (web search falls back to DuckDuckGo). Set OPENAI_API_KEY for LLM categorization and TAVILY_API_KEY / YOUTUBE_API_KEY for higher-quality discovery. See .env.example.

Tech stack

• Backend: FastAPI, SQLModel, SQLite (dev) / Postgres / Supabase (prod)
• Frontend: React + Vite + TypeScript
• Discovery: DuckDuckGo / Tavily web search, YouTube Data API
• Classification: OpenAI-compatible LLM (structured output)
• Deploy: Cloudflare Workers/Pages (frontend + proxy), Render / Cloud Run / Docker (backend)

Data & privacy

This repository ships with synthetic sample data only (creator-discovery/data/sample_creators.csv).

It does not include any real creator dataset. Files that may contain real contact information — *.xlsx source workbooks and generated influencers.json snapshots — are git-ignored and must never be committed. Populate your own database with the seed/import scripts under creator-discovery/scripts/.

Contact emails are private by default. Scraped contact emails are stripped from every API response and export unless you opt in with EXPOSE_CONTACT_EMAILS=true. Keep this false on any public deployment; enable it only on a private/trusted instance where you need emails for outreach. This means a public deployment can browse the full catalog (handles, niches, follower counts, etc.) served from your backend without exposing anyone's email address.

If you use this tool to collect public profile data, make sure you comply with each platform's Terms of Service and applicable privacy laws (GDPR/CCPA, etc.).

Contributing

Issues and pull requests are welcome. Please:

1. Keep secrets out of commits — use .env (git-ignored), never hardcode API keys.
2. Never commit real creator/contact datasets.
3. Run pytest (in creator-discovery/) before opening a PR.

License

MIT © 2026 TheOnlyJason