Jan - Configuration

Environment Variables

Jan Server configuration is managed through environment variables defined in the Helm values file at charts/umbrella-chart/values.yaml.

API Gateway Configuration

Core Settings

Variable	Default	Description
`JAN_INFERENCE_MODEL_URL`	`http://jan-server-jan-inference-model:8101`	Internal URL for inference service

Authentication

Variable	Purpose	Format
`JWT_SECRET`	JWT token signing	Base64 encoded HMAC-SHA256 key
`APIKEY_SECRET`	API key signing	Base64 encoded HMAC-SHA256 key

The default JWT and API key secrets are for development only. Generate new secrets for production deployments.

OAuth2 Integration

Variable	Description
`OAUTH2_GOOGLE_CLIENT_ID`	Google OAuth2 application client ID
`OAUTH2_GOOGLE_CLIENT_SECRET`	Google OAuth2 application secret
`OAUTH2_GOOGLE_REDIRECT_URL`	Callback URL for OAuth2 flow

External APIs

Variable	Provider	Purpose
`SERPER_API_KEY`	Serper	Web search integration

Database Connection

Variable	Default	Description
`DB_POSTGRESQL_WRITE_DSN`	`host=jan-server-postgresql user=jan-user password=jan-password dbname=jan port=5432 sslmode=disable`	Write database connection
`DB_POSTGRESQL_READ1_DSN`	`host=jan-server-postgresql user=jan-user password=jan-password dbname=jan port=5432 sslmode=disable`	Read database connection

Helm Configuration

Updating Values

Edit the configuration in charts/umbrella-chart/values.yaml:


jan-api-gateway:
  env:
    - name: SERPER_API_KEY
      value: your_serper_api_key
    - name: OAUTH2_GOOGLE_CLIENT_ID  
      value: your_google_client_id
    - name: OAUTH2_GOOGLE_CLIENT_SECRET
      value: your_google_client_secret

Applying Changes

After modifying values, redeploy the application:


helm upgrade jan-server ./charts/umbrella-chart

Authentication Setup

JWT Tokens

Generate a secure JWT signing key:


# Generate 256-bit key for HMAC-SHA256
openssl rand -base64 32

Update the JWT_SECRET value in your Helm configuration.

API Keys

Generate a secure API key signing secret:


# Generate 256-bit key for HMAC-SHA256
openssl rand -base64 32

Update the APIKEY_SECRET value in your Helm configuration.

Google OAuth2

Create Google Cloud Project
- Go to Google Cloud Console (opens in a new tab)
- Create a new project or select existing
Enable OAuth2
- Navigate to "APIs & Services" > "Credentials"
- Create OAuth2 client ID credentials
- Set application type to "Web application"
Configure Redirect URI

http://localhost:8080/auth/google/callback
Update Configuration
- Set OAUTH2_GOOGLE_CLIENT_ID to your client ID
- Set OAUTH2_GOOGLE_CLIENT_SECRET to your client secret
- Set OAUTH2_GOOGLE_REDIRECT_URL to your callback URL

External Integrations

Serper API

Jan Server integrates with Serper for web search capabilities.

Get API Key
- Register at serper.dev (opens in a new tab)
- Generate API key from dashboard
Configure
- Set SERPER_API_KEY in Helm values
- Redeploy the application

Adding New Integrations

To add new external API integrations:

Update Helm Values

jan-api-gateway: env: - name: YOUR_API_KEY value: your_api_key_value
Update Go Configuration

Add to config/environment_variables/env.go:

YourAPIKey string `env:"YOUR_API_KEY"`
Redeploy

helm upgrade jan-server ./charts/umbrella-chart

Database Configuration

Connection Settings

The default PostgreSQL configuration uses:

Host: jan-server-postgresql (Kubernetes service name)
Database: jan
User: jan-user
Password: jan-password
Port: 5432
SSL: Disabled (development only)

Production Database

For production deployments:

External Database
- Use managed PostgreSQL service (AWS RDS, Google Cloud SQL)
- Update DSN variables with external connection details
SSL/TLS
- Enable sslmode=require in connection strings
- Configure certificate validation
Connection Pooling
- Consider using connection pooler (PgBouncer, pgpool-II)
- Configure appropriate pool sizes

Model Configuration

The inference model service is configured via Docker CMD parameters:


CMD ["--model", "/models/Jan-v1-4B", \
     "--served-model-name", "jan-v1-4b", \
     "--host", "0.0.0.0", \
     "--port", "8101", \
     "--max-num-batched-tokens", "1024", \
     "--enable-auto-tool-choice", \
     "--tool-call-parser", "hermes", \
     "--reasoning-parser", "qwen3"]

Model Parameters

Parameter	Value	Description
`--model`	`/models/Jan-v1-4B`	Path to model files
`--served-model-name`	`jan-v1-4b`	API model identifier
`--max-num-batched-tokens`	`1024`	Maximum tokens per batch
`--tool-call-parser`	`hermes`	Tool calling format
`--reasoning-parser`	`qwen3`	Reasoning output format

Model configuration changes require rebuilding the inference Docker image. This will be configurable via environment variables in future releases.

Resource Configuration

Kubernetes Resources

Current deployments use default resource limits. For production:


jan-api-gateway:
  resources:
    requests:
      cpu: 100m
      memory: 128Mi
    limits:
      cpu: 500m
      memory: 512Mi
jan-inference-model:
  resources:
    requests:
      cpu: 1000m
      memory: 4Gi
    limits:
      cpu: 4000m
      memory: 8Gi

Storage

PostgreSQL uses default Kubernetes storage. For production:


postgresql:
  persistence:
    enabled: true
    size: 20Gi
    storageClass: fast-ssd

Logging Configuration

Configure logging levels via environment variables:


jan-api-gateway:
  env:
    - name: LOG_LEVEL
      value: info
    - name: LOG_FORMAT
      value: json

Available log levels: debug, info, warn, error Available formats: text, json

Installation API Reference