prshant28/x247_gen_ai

---

✨ What is Recall X247?

Recall X247 is a production-ready multi-agent AI system built for the Gen AI Academy APAC 2026 Hackathon. It acts as your intelligent Second Brain — a unified platform that:

• 📥 Captures knowledge from YouTube videos, web articles, PDFs, and plain notes
• 🔍 Recalls information using natural language semantic search with AI-powered answers
• ✅ Manages tasks and projects linked directly to your captured knowledge
• 📅 Schedules study sessions and events on your calendar
• 🃏 Generates flashcards for active recall and spaced repetition
• 📊 Delivers AI daily briefings and learning analytics
• 🤖 Orchestrates all of this through a single natural language interface

---

🚀 Live Demo

---

🏗️ Multi-Agent Architecture

Recall X247 uses a hierarchical multi-agent system with an Orchestrator coordinating 6 specialized sub-agents via OpenAI function calling with real-time SSE streaming:

User Natural Language Request
         │
         ▼
┌─────────────────────────────────────┐
│   🧠 Orchestrator (KnowledgeCoordinator)  │
│   GPT-4o-mini via OpenRouter        │
│   app/coordinator.py                │
└──────────────┬──────────────────────┘
               │ OpenAI Function Calling
               │ (10 MCP-style tools)
    ┌──────────┼──────────────────┐
    │          │                  │
    ▼          ▼                  ▼
┌────────┐ ┌────────┐      ┌──────────┐
│Capture │ │Recall  │      │  Task    │
│ Agent  │ │ Agent  │      │  Agent   │
│📥     │ │🔍     │      │ ✅      │
└────────┘ └────────┘      └──────────┘
    │          │                  │
    ▼          ▼                  ▼
┌──────────┐ ┌────────────┐ ┌──────────────┐
│Calendar  │ │ Briefing   │ │  Analytics   │
│  Agent   │ │   Agent    │ │    Agent     │
│ 📅      │ │ 📊        │ │  📈         │
└──────────┘ └────────────┘ └──────────────┘
         │
         ▼ SSE Streaming (real-time events)
┌─────────────────────────────────┐
│   Frontend Agent Hub (/agent)   │
│   React + TypeScript + Vite     │
└─────────────────────────────────┘

Agent Registry

---

🎨 Features

---

🛠️ Tech Stack

---

📂 Project Structure

Gen_AI_APAC/
├── 📄 main.py                    # FastAPI entry point
├── 📄 index.html                 # HTML shell
├── 📄 Dockerfile                 # Container config (Cloud Run)
├── 📄 vite.config.ts             # Vite + proxy config
├── 📄 requirements.txt           # Python dependencies
├── 📄 package.json               # Node dependencies
│
├── app/                          # Python backend
│   ├── config.py                 # Env vars & OpenRouter config
│   ├── db.py                     # Firestore + in-memory fallback
│   ├── coordinator.py            # Orchestrator (10 tools, SSE streaming)
│   ├── workflow_engine.py        # Workflow tracking + AGENT_REGISTRY
│   ├── capture_agent.py          # YouTube / web / PDF / note ingestion
│   ├── recall_agent.py           # 3-tier semantic search
│   ├── task_agent.py             # Task CRUD
│   └── calendar_agent.py         # Google Calendar integration
│
├── src/                          # React frontend
│   ├── App.tsx                   # All 13 views + SSE streaming
│   └── index.css                 # Dark neural theme + animations
│
└── .github/workflows/
    ├── deploy.yml                # Cloud Run CI/CD
    └── gh-pages.yml              # GitHub Pages CI/CD

---

⚙️ Quick Start

Prerequisites

• Python 3.11+
• Node.js 20+
• An OpenRouter API key (GEN_APAC_API_KEY)

1. Clone & Install

git clone https://github.com/prshant28/Gen_AI_APAC.git
cd Gen_AI_APAC

# Install Python dependencies
pip install -r requirements.txt

# Install Node dependencies
npm install

2. Configure Environment

cp .env.example .env

Edit .env with your keys:

# Required
GEN_APAC_API_KEY=your_openrouter_api_key

# Optional — for full Firestore persistence
GCP_PROJECT_ID=your_gcp_project_id
GOOGLE_APPLICATION_CREDENTIALS=path/to/service_account.json

# Optional — for Google Custom Search
GOOGLE_CSE_CX=your_cse_id
GOOGLE_API_KEY=your_google_api_key

💡 No Firestore? The app uses an in-memory mock database automatically — just set GEN_APAC_API_KEY and you’re ready to go!

3. Run Locally

npm run dev

This starts both the FastAPI backend (port 8000) and Vite frontend (port 5000) concurrently. Open http://localhost:5000.

4. Docker (Optional)

docker build -t recall-x247 .
docker run -p 8000:8000 --env-file .env recall-x247

---

📡 API Reference

Core Endpoints

Chat API Example

curl -X POST https://recall-x247-asia-southeast1.run.app/chat \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Capture this YouTube video https://youtu.be/dQw4w9WgXcQ and create a task to review it",
    "user_id": "demo_user"
  }'

SSE Streaming Events

The /agent/chat/stream endpoint emits real-time events:

---

🖥️ Application Views (13 Total)

---

🚢 Deployment

Automated CI/CD

Every push to main triggers two parallel deployments:

Push to main
    ├── 🐳 Cloud Run (deploy.yml)
    │   ├── Build Docker image
    │   ├── Deploy to asia-southeast1
    │   └── Health check /health & /api/health
    │
    └── 📄 GitHub Pages (gh-pages.yml)
        ├── npm run build
        └── Deploy to GitHub Pages

Manual Cloud Run Deployment

gcloud run deploy recall-x247 \
  --source . \
  --region asia-southeast1 \
  --allow-unauthenticated \
  --memory 1Gi \
  --cpu 1 \
  --set-secrets "GEN_APAC_API_KEY=GEN_APAC_API_KEY:latest"

Required GitHub Secrets

---

🛡️ Security

• Data Isolation: Firestore security rules ensure users can only access their own data (see firestore.rules)
• Secret Management: API keys stored as Google Cloud Secrets, never in source code
• Authentication: Cloud Run requires valid credentials for write operations
• CORS: FastAPI CORS middleware configured for production domains

---

🤝 Contributing

1. Fork the repository
2. Create your feature branch: git checkout -b feat/amazing-feature
3. Commit your changes: git commit -m 'feat: add amazing feature'
4. Push to the branch: git push origin feat/amazing-feature
5. Open a Pull Request

---

Live Demo

Example Prompts:

• Prompt 1: What are the benefits of AI?
• Prompt 2: How can AI change the world?

Watch our live demo: YouTube Video

📜 License

This project is licensed under the MIT License.

---