# LLM Proxy Environment Variables
# Copy this file to .env and edit with your settings

# =============================================================================
# SERVER CONFIGURATION
# =============================================================================
# Port to listen on (default: 9999)
PORT=9999

# Memory threshold for the pool in GB (default: 16GB)
MEMORY_THRESHOLD_GB=16

# =============================================================================
# RATE LIMITING (Make Optional)
# =============================================================================
# Maximum tokens allowed per request window (only when enabled)
# Set to empty or use DISABLE_RATE_LIMITING=true to remove rate limiting
RATE_LIMIT_MAX_TOKENS=100

# Tokens per second refill rate for rate limiter (default: 10/sec)
RATE_LIMIT_REFILL_RATE=10

# Completely disable rate limiting for maximum throughput (default: false)
# Set to true if you don't need rate limiting
DISABLE_RATE_LIMITING=false

# =============================================================================
# MODEL DISCOVERY (LM Studio Integration)
# =============================================================================
# Enable discovery from LM Studio (default: true)
DISCOVERY_ENABLED=true

# URL for LM Studio discovery endpoint (default: localhost:1234)
LM_STUDIO_DISCOVERY_URL=http://localhost:1234/api/v1/models

# =============================================================================
# API AUTHENTICATION (Optional)
# =============================================================================
# Set API keys for models if using authentication
# Leave empty if not using auth
LLM_API_KEY_QWEN=your-qwen-api-key-here
LLM_API_KEY_MISTRAL=your-mistral-api-key-here

# =============================================================================
# LOGGING
# =============================================================================
# Log level: debug, info, warning, error (default: info)
LOG_LEVEL=info