enVector - Encrypted Vector Search

enVector is a product that provides secure vector search functionality using ES2 (Encrypted Similarity Search), which is based on Fully Homomorphic Encryption (FHE). This repository contains self-hosted deployment scripts and client SDK examples.

🚀 Features

• 🔐 End-to-End Encryption: Secure vector search with FHE (Fully Homomorphic Encryption)
• ⚡ High Performance: Optimized vector similarity search algorithms
• 🏗️ Microservices Architecture: Scalable and maintainable service design
• 🐳 Multi-Platform Deployment: Docker Compose and Kubernetes (Helm) support
• 📱 Python SDK: Easy-to-use client library for integration

🔒 Security Features

• Fully Homomorphic Encryption: Secure computation on encrypted data
• Client-side Key Management: Secret key never leave the client
• Encrypted Vector Storage: All vector data is encrypted at rest (at-rest)
• Secure Search: Search operations performed on encrypted data (in-use)

📊 Performance

• Vector Dimensions: Support for 16-4096 dimensional vectors
• Search Speed: Optimized encrtyped similarity search algorithms
• Scalability: Horizontal scaling with Kubernetes
• GPU Support: Optional GPU acceleration for encrypted vector search

🛠️ Development

Prerequisites

• Python 3.9-3.13
• Docker and Docker Compose
• Kubernetes cluster (for K8s deployment)
• Helm 3.0+
• Linux or macOS 11.0+

🏗️ Architecture

enVector consists of four main microservices:

• es2e (Endpoint): Main API gateway and client interface
• es2b (Backend): Service orchestration and metadata management
• es2s (Search): Vector search engine and similarity computation
• es2c (Compute): Vector operations and mathematical computations

Infrastructure Dependencies

• PostgreSQL: Metadata storage and management
• MinIO: Vector data storage (S3-compatible)

📁 Project Structure

envector-deployment/
├── docker-compose/          # Docker Compose deployment
│   ├── docker-compose.yml   # Multi-service orchestration
│   └── README.md           # Docker setup guide
├── kubernetes-manifests/    # Kubernetes deployment
│   ├── helm/               # Helm chart for K8s
│   │   ├── Chart.yaml      # Chart metadata
│   │   ├── values.yaml     # Configurable values
│   │   └── templates/      # K8s manifest templates
│   └── README.md           # K8s deployment guide
└── notebooks/              # Python SDK examples    

🚀 Quick Start

⚠️ Important: Docker images are stored in private repositories. Please contact heaan for access credentials before proceeding with deployment.

Option 1: Docker Compose

Recommended for Development.

Method A: Clone Repository

# Clone the repository
git clone https://github.com/CryptoLabInc/envector-deployment.git
cd envector-deployment

# Copy environment file
cp .env.example .env

# Start services
docker compose -f docker-compose/docker-compose.yml -p envector up -d

Method B: Direct HTTP Usage

# Download and run directly from GitHub
curl -O https://raw.githubusercontent.com/cryptolabinc/envector-deployment/main/docker-compose/docker-compose.yml
curl -O https://raw.githubusercontent.com/cryptolabinc/envector-deployment/main/docker-compose/.env.example

# Copy environment file
cp .env.example .env

# Start services
docker compose -f docker-compose.yml -p envector up -d

Option 2: Kubernetes

Recommended for production.

Clone Repository

# Clone the repository
git clone https://github.com/CryptoLabInc/envector-deployment.git
cd envector-deployment

# Install Helm chart
helm install envector ./kubernetes-manifests/helm

# Check deployment status
kubectl get pods

# Access services
kubectl get svc

🔧 Configuration

Environment Variables

Variable	Description	Default
ES2E_TAG	es2e service image tag	latest
ES2B_TAG	es2b service image tag	latest
ES2S_TAG	es2s service image tag	latest
ES2C_TAG	es2c service image tag	latest
ES2_LOG_LEVEL	Logging level	INFO
ES2E_HOST_PORT	es2e external port	50050

Helm Values

Edit kubernetes-manifests/helm/values.yaml to customize:

• Service ports and types
• Resource limits and replicas
• External database connections
• Image repositories and tags

📚 Python SDK Usage

• Python Version: 3.9-3.13
• OS: Linux/macOS 11.0+

pip install es2

Basic Setup

import es2

# Initialize connection
es2.init(
    host="localhost",
    port=50050,
    key_path="./keys",
    key_id="my_key"
)

# Create index
index = es2.create_index("my_index", dim=512)

# Insert vectors
vectors = [...]  # Your 512-dimensional vectors
index.insert(vectors, metadata=["doc1", "doc2"])

# Search
results = index.search(query_vector, top_k=5)

Key Management

from es2.crypto import KeyGenerator, Cipher

# Generate FHE keys
keygen = KeyGenerator("./keys/my_key")
keygen.generate_keys()

# Create cipher for encryption/decryption
cipher = Cipher(dim=512, enc_key_path="./keys/my_key/EncKey.bin")

📄 License

This project is proprietary software. For licensing information, please contact heaan.

🤝 Contributing

This is a proprietary software project. For contribution inquiries, please contact heaan.

📞 Support

• Product Information: enVector at heaan
• Technical Support: Please contact heaan

🔗 Related Links

• enVector Product Page
• heaan
• FHE Resources

---