Skip to main content

​
Environment Variables

MCPHub uses environment variables for configuration. This guide covers all available variables and their usage.

​
Core Application Settings

​
Server Configuration

VariableDefaultDescription
PORT3000Port number for the HTTP server
INIT_TIMEOUT300000Initial timeout for the application
BASE_PATH''The base path of the application
READONLYfalseSet to true to enable readonly mode
MCPHUB_SETTING_PATHPath to the MCPHub settings
NODE_ENVdevelopmentApplication environment (development, production, test)
PORT=3000
INIT_TIMEOUT=300000
BASE_PATH=/api
READONLY=true
MCPHUB_SETTING_PATH=/path/to/settings
NODE_ENV=production

​
Authentication & Security

​
Admin Password

VariableDefaultDescription
ADMIN_PASSWORD(random)Password for the default admin user. If not set, a cryptographically random 24-character password is generated on first launch and printed to the server logs.
ADMIN_PASSWORD=your-secure-admin-password

​
JWT Configuration

VariableDefaultDescription
JWT_SECRET-Secret key for JWT token signing (required)
JWT_SECRET=your-super-secret-key-change-this-in-production

​
Smart Routing & Embeddings

​
Embedding Configuration

VariableTypeDefaultDescription
SMART_ROUTING_ENABLEDbooleanfalseEnable the smart routing feature
EMBEDDING_MODELstringtext-embedding-3-smallEmbedding model name used for vector search
AZURE_OPENAI_EMBEDDING_MODELstringtext-embedding-3-smallThe actual OpenAI model name deployed in Azure (e.g. text-embedding-3-small). Azure deployment names are arbitrary identifiers, so this field tells MCPHub which OpenAI model is behind the deployment, enabling correct token-limit enforcement and tokenizer selection. Only used when SMART_ROUTING_EMBEDDING_PROVIDER=azure_openai.
EMBEDDING_MAX_TOKENSnumberBy modelMaximum tokens for text truncation before generating embeddings. Overrides the per-model default. Useful to match the batch_size of a local inference server (e.g. LocalAI, LM Studio). When unset, the limit is resolved automatically: text-embedding-* → 8191, bge-m3 → 8192, gemini-embedding-001 → 2048, other BGE models → 512, unknown models → 512.
# Limit truncation to 510 tokens to match a LocalAI batch_size=512 setup
EMBEDDING_MAX_TOKENS=510

​
Configuration Examples

​
Development Environment

# .env.development
NODE_ENV=development
PORT=3000

# Auth
JWT_SECRET=dev-secret-key

​
Production Environment

# .env.production
NODE_ENV=production
PORT=3000

# Security
JWT_SECRET=your-super-secure-production-secret

​
Docker Environment

# .env.docker
NODE_ENV=production
PORT=3000

# Security
JWT_SECRET_FILE=/run/secrets/jwt_secret

​
Environment Variable Loading

MCPHub loads environment variables in the following order:
  1. System environment variables
  2. .env.local (ignored by git)
  3. .env.{NODE_ENV} (e.g., .env.production)
  4. .env

​
Using dotenv-expand

MCPHub supports variable expansion: 
BASE_URL=https://api.example.com
API_ENDPOINT=${BASE_URL}/v1

​
Security Best Practices

  1. Never commit secrets to version control
  2. Use strong, unique secrets for production
  3. Rotate secrets regularly
  4. Use environment-specific files
  5. Validate all environment variables at startup
  6. Use Docker secrets for container deployments