🧠 SecondBrain — Personal AI Life OS

AI-powered personal productivity OS with RAG-based chat, task management, and semantic notes. Built with Angular, Spring Boot, PostgreSQL, and Google Gemini.

🌐 Live Demo → https://d3icm4v3pyidtw.cloudfront.net

---

✨ What is SecondBrain?

SecondBrain is a personal knowledge and task management system with a built-in AI reasoning layer. Instead of a simple chatbot, it uses Retrieval-Augmented Generation (RAG) to give the AI real context about your tasks and notes — so it can actually help you think, not just respond.

Ask it things like:

• "What are my high priority tasks this week?"
• "Summarize my note about Project X"
• "Create a task to call John tomorrow at 5 PM"

---

🚀 Features

🤖 AI Assistant (Gemini 2.5 Flash Lite)

• Real-time streaming responses via Server-Sent Events (SSE)
• RAG pipeline — AI reads your actual tasks and notes before answering
• Decision Engine — routes queries intelligently (AI / Hybrid / Direct DB)
• Semantic search using vector embeddings stored in PostgreSQL
• Persistent chat history that survives server restarts
• Agentic actions — AI can create tasks directly from chat (with confirmation)
• Time-aware greetings (Good morning / afternoon / evening / night)

✅ Task Management

• Create, edit, delete tasks
• Priority levels — Low / Medium / High with color badges
• Due dates with overdue highlighting
• Status lifecycle — Todo → In Progress → Done
• "Active today" count

📝 Knowledge Base (Notes)

• Rich note editor with title + content
• AI Assistant sidebar per note — Summarize, Explain, Extract actions, Improve writing
• Semantic search across all notes
• Auto-embedding on save for RAG retrieval

🔐 Authentication

• JWT-based auth with secure token storage
• Auto-login after registration
• Per-user data isolation — users only see their own data
• Rate limiting on auth endpoints

🎨 UI/UX

• Premium Glassmorphism dark theme
• Light / Dark mode toggle with OS preference detection
• Responsive layout (desktop + tablet)
• Real-time streaming text with typing cursor animation
• Suggested prompt chips for new users

---

🛠️ Tech Stack

Layer	Technology
Frontend	Angular 12, TypeScript, Vanilla CSS
Backend	Spring Boot 3.4, Java 17
Database	PostgreSQL (+ vector embeddings)
AI	Google Gemini 2.5 Flash Lite
Auth	JWT (JSON Web Tokens)
Streaming	Server-Sent Events (SSE)
Embeddings	Gemini Embedding API + DoubleListConverter
CI/CD	GitHub Actions, Docker, AWS ECR
Infrastructure	AWS (EC2, S3, CloudFront, RDS)
Build	Maven, Angular CLI

---

📸 Screenshots

Login	Tasks

Notes	Note Detail + AI

AI Chat	AI Task Creation

---

🏗️ Project Structure

secondBrain/
├── backend/                          # Spring Boot application
│   └── src/main/java/com/example/backend/
│       ├── ai/                       # AI orchestration layer
│       │   ├── AiChatService.java    # Main AI pipeline
│       │   ├── DecisionEngineService.java  # Intent routing
│       │   ├── ContextBuilderService.java  # RAG context assembly
│       │   ├── GeminiService.java    # Gemini API client (SSE)
│       │   ├── EmbeddingService.java # Vector embeddings
│       │   ├── SafetyService.java    # Input sanitization
│       │   ├── SemanticCacheService.java   # Response caching
│       │   └── PromptBuilder.java    # System prompt construction
│       ├── controller/               # REST API endpoints
│       ├── service/                  # Business logic
│       ├── model/                    # JPA entities
│       ├── repository/               # Spring Data repositories
│       ├── security/                 # JWT auth + rate limiting
│       └── dto/                      # Data transfer objects
│
└── frontend/                         # Angular application
    └── src/app/
        ├── features/
        │   ├── ai-chat/              # AI chat page + streaming
        │   ├── tasks/                # Task management
        │   ├── notes/                # Notes list + detail editor
        │   └── auth/                 # Login + Register
        └── core/
            ├── guards/               # Auth route guards
            └── interceptors/         # JWT HTTP interceptor

---

☁️ Deployment Architecture

The application is deployed on AWS (ap-south-1 — Mumbai) with an automated CI/CD pipeline for the backend and manual deployment for the frontend.

🔗 Live at: https://d3icm4v3pyidtw.cloudfront.net

AWS Services Used

Service	Region	Role
S3	ap-south-1	Hosts the Angular SPA (static build files)
CloudFront	Global (Edge)	CDN — serves frontend over HTTPS, proxies /api/* requests to backend
EC2	ap-south-1	Runs the Dockerized Spring Boot backend
RDS (PostgreSQL)	ap-south-1	AWS-managed relational database with vector embedding storage
ECR	ap-south-1	Private Docker image registry
SSM	ap-south-1	Triggers container restart on deploy — no SSH required

Backend — CI/CD Pipeline (Automated)

1. Developer pushes code to main branch on GitHub
2. GitHub Actions automatically:
   • Sets up JDK 17 (Temurin)
   • Builds the Spring Boot JAR via Maven
   • Builds a multi-stage Docker image (Maven build → lightweight JRE runtime — keeps the final image small and secure)
   • Pushes the image to AWS ECR
   • Injects all secrets (Gemini API key, DB credentials, CORS origins) from GitHub Secrets — nothing hardcoded
3. AWS SSM sends a remote command to the EC2 instance to:
   • Pull the latest image from ECR
   • Stop and replace the running container
   • Start the new container with production environment variables
   • Prune old images to save disk space
4. Spring Boot (in Docker on EC2) handles business logic, JWT auth, and communicates with RDS PostgreSQL for data and Google Gemini API for AI

Frontend — Manual Deployment

The Angular frontend is built locally and synced to S3:

cd frontend/secondbrain-frontend
ng build --configuration production
aws s3 sync dist/secondbrain-frontend s3://your-s3-bucket --delete

CloudFront then serves the SPA globally over HTTPS and routes /api/* to the EC2 backend origin.

Docker Setup

The backend uses a multi-stage Dockerfile to keep the production image lean:

# Build stage — full Maven + JDK
FROM maven:3.9.6-eclipse-temurin-17 AS build
WORKDIR /app
COPY . .
RUN mvn clean package -DskipTests

# Run stage — JRE only (no build tools)
FROM eclipse-temurin:17-jre
COPY --from=build /app/target/*.jar app.jar
ENTRYPOINT ["java", "-jar", "app.jar"]

This produces a minimal runtime image without Maven, source code, or build artifacts — reducing image size and attack surface.

Automated deployment with minimal downtime — push to main and the backend is live in ~2 minutes.

---

⚙️ Getting Started

Prerequisites

• Java 17+
• Node.js 18+
• PostgreSQL 14+
• Google Gemini API key (Get one free here)

1. Clone the repo

git clone https://github.com/tanmayythakare/Smart-SecondBrain.git
cd secondbrain

2. Set up the database

CREATE DATABASE secondbrain_db;
CREATE USER secondbrain_user WITH PASSWORD 'your_password';
GRANT ALL PRIVILEGES ON DATABASE secondbrain_db TO secondbrain_user;

3. Configure the backend

Edit backend/src/main/resources/application.properties or set the equivalent environment variables:

spring.datasource.url=jdbc:postgresql://localhost:5432/secondbrain_db
spring.datasource.username=secondbrain_user
spring.datasource.password=your_password

jwt.secret=your_jwt_secret_key_here

app.ai.gemini.key=your_gemini_api_key_here
app.ai.gemini.model=gemini-2.5-flash-lite
app.ai.gemini.endpoint=https://generativelanguage.googleapis.com/v1beta/models/

Environment Variables Reference

Variable	Required	Description
spring.datasource.url	✅	JDBC connection string for PostgreSQL
spring.datasource.username	✅	Database username
spring.datasource.password	✅	Database password
jwt.secret	✅	Secret key for signing JWT tokens (min 32 chars recommended)
app.ai.gemini.key	✅	Google Gemini API key
app.ai.gemini.model	❌	Gemini model name (default: gemini-2.5-flash-lite)
app.ai.gemini.endpoint	❌	Gemini API base URL (default: Google's public endpoint)

⚠️ Never commit API keys or secrets to version control. Use environment variables or a .env file excluded via .gitignore.

4. Run the backend

cd backend
./mvnw spring-boot:run

Backend runs on http://localhost:8080

5. Run the frontend

cd frontend/secondbrain-frontend
npm install
ng serve

Frontend runs on http://localhost:4200

6. Open the app

Visit http://localhost:4200 — register an account and start using SecondBrain.

---

🔌 API Endpoints

Method	Endpoint	Description
POST	/api/auth/register	Register new user
POST	/api/auth/login	Login, returns JWT
GET	/api/tasks	Get all tasks
POST	/api/tasks	Create task
PUT	/api/tasks/{id}	Update task
DELETE	/api/tasks/{id}	Delete task
GET	/api/notes	Get all notes
POST	/api/notes	Create note
PUT	/api/notes/{id}	Update note
DELETE	/api/notes/{id}	Delete note
GET	/api/notes/search?q=	Semantic search notes
POST	/api/ai/chat/stream	AI chat (SSE stream)
GET	/api/ai/chat/history	Get chat history
POST	/api/ai/chat/confirm	Execute agentic action

All endpoints except /api/auth/** require a valid JWT in the Authorization: Bearer <token> header.

---

🧠 How the AI Works

User Message
     │
     ▼
SafetyService ──── (blocks harmful input)
     │
     ▼
DecisionEngine ─── (AI / Hybrid / Direct DB / Action)
     │
     ├── NON_AI ──► Direct DB query → instant response
     │
     ├── ACTION ──► Extract payload → Confirm button → Execute
     │
     └── AI/HYBRID ──► ContextBuilder (RAG)
                              │
                        ┌─────┴──────┐
                     Recent      Semantic
                     Items       Search
                        └─────┬──────┘
                              │
                         PromptBuilder
                              │
                         GeminiService
                         (SSE Stream)
                              │
                         Frontend renders
                         token by token

Key concepts:

• Decision Engine classifies each user message into one of four routes — avoiding unnecessary AI calls for simple queries
• RAG (Retrieval-Augmented Generation) pulls relevant tasks and notes into the prompt context so the AI has real, personal data to reason about
• Agentic Actions let the AI propose creating tasks, with a confirmation step before execution — keeping the user in control
• SSE Streaming delivers tokens in real-time for a responsive, ChatGPT-like experience

---

🔒 Security

• All endpoints (except /api/auth/**) require a valid JWT
• Passwords hashed with BCrypt
• Rate limiting on auth and AI endpoints
• Input sanitization before AI processing
• Per-user data isolation enforced at service layer
• Environment variables for all secrets — no hardcoded keys in source code
• CORS configured for frontend origin only

---

❓ Troubleshooting

CORS errors in browser

Ensure the backend's CORS configuration allows your frontend origin (http://localhost:4200 for local dev). The SecurityConfig.java permits OPTIONS preflight requests — if you've changed ports, update accordingly.

AI chat returns errors or empty responses

1. Verify your app.ai.gemini.key is valid at Google AI Studio
2. Check backend logs for 403 or 429 (quota exceeded) responses from the Gemini API
3. Ensure the model name in config matches an available model

Frontend 404 on page refresh (production)

The Angular app uses client-side routing. In production (S3/CloudFront), configure the error document to redirect to index.html. For hash-based routing, the app already uses useHash: true.

Database connection refused

1. Confirm PostgreSQL is running: pg_isready -h localhost -p 5432
2. Verify the database and user exist with the credentials in your config
3. Check that pg_hba.conf allows local connections

---

🚧 Future Improvements

Things I'm planning to add next:

• 📱 Progressive Web App (PWA) support for mobile
• 🎙️ Voice input for AI chat
• 🏷️ Tags and categories for tasks and notes
• 📊 Dashboard with productivity analytics
• 🔗 Multi-model support (GPT-4o, Claude, Llama)
• 📤 Export notes as PDF / Markdown
• 🔔 Push notifications for due tasks
• 🐳 Docker Compose for one-command local setup
• 🧪 Unit and integration test coverage
• 🌐 Automated frontend CI/CD via GitHub Actions

---

📋 What I Learned Building This

This project was built to learn and demonstrate:

• Full-stack development with Angular + Spring Boot
• AI integration — not just calling an API, but building a proper RAG pipeline
• Streaming UIs — Server-Sent Events for real-time token streaming
• Vector embeddings — storing and querying semantic vectors in PostgreSQL
• JWT authentication — stateless auth with Spring Security
• Cloud deployment — Dockerized backend on EC2, Angular SPA on S3 + CloudFront
• CI/CD pipelines — GitHub Actions for automated build, test, and deploy
• Production practices — error handling, rate limiting, data isolation, environment config

---

🤝 Contributing

Contributions are welcome! If you'd like to improve SecondBrain:

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

---

👤 Author

Tanmay Thakare

• GitHub: @tanmayythakare

---

📄 License

This project is open source and available under the MIT License.

---

Built with ☕ and a lot of debugging