FlowBoard

Direct Your Video Frame by Frame

FlowBoard.Demo.mp4

AI-powered video storyboarding that transforms sketches into context-aware video clips. Draw, prompt, generate—infinitely.

---

✨ Features

Feature	Description
🎨 Interactive Canvas	Draw instructions directly on frames using Tldraw
🤖 AI Video Generation	Powered by Google Vertex AI (Veo 3.1 & Gemini 2.5)
🔗 Frame-by-Frame Workflow	Sequential frames connected by arrows build your story
⚡ Image Enhancement	AI-powered frame improvement on demand
🎬 Video Merging	Combine clips into seamless sequences

---

🚀 Local Setup

Prerequisites

• Python 3.11+ & Node.js 18+
• Redis (local or hosted)
• Google Cloud Project (Vertex AI enabled)
• Supabase project

Backend Setup

cd backend
pip install -r requirements.txt

.env Configuration:

# Google Cloud / Vertex AI
GOOGLE_CLOUD_PROJECT=your-gcp-project-id
GOOGLE_CLOUD_LOCATION=us-central1
GOOGLE_GENAI_USE_VERTEXAI=true
GOOGLE_CLOUD_BUCKET_NAME=your-gcs-bucket-name

# Redis (job queue)
REDIS_URL=redis://default:password@localhost:6379

# Supabase (auth & database)
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_SECRET_KEY=your-service-role-key

Setup Checklist:

• ✅ Enable Vertex AI API + create GCS bucket
• ✅ Auth: GOOGLE_APPLICATION_CREDENTIALS or gcloud auth application-default login
• ✅ Supabase: Create users table with credits column (see backend/scripts/db)
• ✅ Enable auth providers (Google/GitHub) in Supabase dashboard

Frontend Setup

cd frontend
npm install

.env Configuration:

VITE_BACKEND_URL=http://localhost:8000
VITE_SUPABASE_URL=https://your-project.supabase.co
VITE_SUPABASE_PUBLIC_KEY=your-anon-public-key  # NOT service role key

Run

Backend: python main.py (→ http://localhost:8000)
Frontend: npm run dev (→ http://localhost:5173)

---

📖 Usage

1. Create Frame → Start with 16:9 frame
2. Draw/Annotate → Sketch motion instructions
3. Add Prompt → Describe your vision
4. Generate → Click ✨ to create next frame
5. Chain Frames → Link frames with arrows for sequences
6. Merge Videos → Combine clips into final output