FlowAI

What it is: An end-to-end AI-powered marketing stack tailored for early-stage startups. With a single platform, you can:

• Turn a basic product image into studio-grade photos via neural rendering.
• Generate polished product videos using just an image.
• Upload CSV contact lists and launch personalized marketing campaigns—voice calls via custom AI assistants and targeted email sequences.
• Plan and strategize using built-in business planner tools.
• Monitor performance with dashboard metrics (emails sent, campaigns run, engagement, video stats, etc.).

---

🔧 Tech Stack

• Frontend: Next.js 15 with TypeScript, Tailwind CSS, Radix UI, React Query, Recharts, Framer Motion

• Backend / API: Next.js API routes (NextAuth w/ Google/YouTube & Prisma with PostgreSQL)

• AI & Media Engines:

  • Google GenAI & GenAI SDKs for prompt-based content
  • Replicate – video generation via veo2
  • Flash 2.5 (experimental) – image generation
  • Bannerbear, Cloudinary, FFMPEG – image/video processing
  • Twilio – voice agents and campaign calls
  • Resend – email delivery

• Data Storage/Processing:

  • PostgreSQL (Prisma & migrations)
  • Redis & Upstash – rate limiting, queues (BullMQ)
  • Qdrant – embeddings & vector similarity
  • CSV parsing, PDF & docx extraction utilities

• Infrastructure: Docker + Docker-Compose for full stack deployment

---

📁 Repository Structure

/actions       → Backend logic for email/video/voice/prompts  
/app/api       → Next.js API routes for auth, content, media, analytics  
/app/*         → Frontend pages and components for each feature (dashboard, video studio, email, voice)  
/lib & prisma  → Shared utilities, db schema, Elasticsearch-type search, embeddings  
/model-voice   → Python service for phone-call voice agents  
public/        → Assets (images, video samples, upload previews)  
Docker setups  → docker-compose.yml and Dockerfile configs for seamless deployment  

---

🎯 Screenshots 🎯

Landing Page

Authentication

Onboarding

Dashboard

Profile

---

🚀 How to Run Locally

1. Clone repo & copy .env.example to .env, fill secrets.
2. docker-compose up --build
3. yarn dev (for local development)
4. Visit http://localhost:3000 — onboard a sample brand, upload a product image, test video & email/voice campaigns, check dashboard metrics.

---

✅ Why It Matters

• Productivity for startups: No need to stitch together multiple tools—FlowAI does it end-to-end with AI at each touchpoint.
• Scalable architecture: Modular actions structure, Next.js APIs, and async processing make adding new channels (e.g., social, chatbots) trivial.
• Future-ready engines: Written to swap in stronger GenAI models or pipelines easily without rewriting frontend.