.env.example

# Gateway Configuration
# This is the only port exposed externally
PORT=8080
DOCS_DIR=./data/docs  # Local folder to mount into the container for ingestion

# LLM Provider Configuration
# Options: openai, ollama, anthropic, gemini, openai_compatible (for LM Studio, Llama.cpp, etc.)
LLM_PROVIDER=ollama
# Recommended for 8GB VRAM: llama3.2 (3b) or llama3 (8b)
# Note: Context window is limited to 8192 in code to prevent OOM
LLM_MODEL=llama3:latest

# Embedding Configuration
EMBEDDING_PROVIDER=ollama
EMBEDDING_MODEL=nomic-embed-text

# Ollama Configuration
# Using the ollama container from docker-compose
OLLAMA_HOST=http://ollama:11434

# OpenAI Compatible Server Configuration (for LM Studio, Llama.cpp, etc.)
# Uncomment and configure these when using LLM_PROVIDER=openai_compatible
# OPENAI_BASE_URL=http://localhost:1234/v1  # Example for LM Studio
# OPENAI_API_KEY=not-needed-for-local  # Many local servers don't require API key

# OpenAI/Anthropic/Gemini Configuration (if applicable)
# OPENAI_API_KEY=your_openai_api_key_here
# ANTHROPIC_API_KEY=your_anthropic_api_key_here
# GOOGLE_API_KEY=your_google_api_key_here

# Ingestion Configuration
INGEST_BATCH_SIZE=10
CHUNK_SIZE=512
CHUNK_OVERLAP=128

# Application Settings
DEBUG=false
LOG_LEVEL=INFO