gabrieltemtsen/airscan

AI-Powered Broadcast Compliance Monitoring Platform for NBC Nigeria # AirScan AirScan is a full-stack SaaS for **broadcast compliance monitoring** for the **National Broadcasting Commission (NBC), Nigeria**. **Pipeline:** Upload → Whisper transcription → Gemini breach analysis → reviewer workflow → PDF/CSV exports. ## Monorepo ``` airscan/ frontend/ # Next.js 14 (App Router) + Clerk backend/ # FastAPI + SQLAlchemy + RQ worker docker-compose.yml .env.example ``` ## Tech Stack - **Frontend:** Next.js 14, TypeScript, Tailwind, shadcn/ui-style components, Clerk auth, Lucide icons, sonner toasts - **Backend:** FastAPI, SQLAlchemy, Postgres - **Jobs:** Redis + RQ - **AI:** OpenAI Whisper (`whisper-1`) → Gemini 2.0 Flash (`gemini-2.0-flash`) - **Storage:** S3-compatible (Cloudflare R2) - **Payments:** Paystack (NGN) ## Local Development ### 1) Start Postgres + Redis ```bash cd airscan docker compose up -d ``` ### 2) Backend setup ```bash cd backend python -m venv .venv source .venv/bin/activate pip install -U pip pip install -e . ``` Create `backend/.env` (copy from repo root `.env.example`) and set values: - `DATABASE_URL` (points to docker postgres) - `REDIS_URL` (points to docker redis) - `CLERK_SECRET_KEY` - `OPENAI_API_KEY` - `GEMINI_API_KEY` - S3/R2 config: `AWS_*` - Paystack: `PAYSTACK_SECRET_KEY` Initialize DB tables and seed the default NBC policy pack: ```bash python -c "from app.main import init_db; init_db()" python seed_policies.py ``` Run the API: ```bash uvicorn app.main:app --reload --port 8000 ``` Run the worker (separate terminal): ```bash python worker.py ``` ### 3) Frontend setup ```bash cd ../frontend npm install ``` Create `frontend/.env.local` with: ```bash NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY= CLERK_SECRET_KEY= NEXT_PUBLIC_API_URL=localhost ``` Run the Next.js app: ```bash npm run dev ``` Visit: - Landing: localhost - Dashboard: localhost ## Key Product Screens - `/` landing page (navy + gold) - `/dashboard` stats + recent cases + usage bar …