nanobanana-mcp-server

# ============================================================
# Authentication Settings
# ============================================================
# NANOBANANA_AUTH_METHOD: Authentication method
#   - api_key: Developer API + API Key (default)
#   - vertex_ai: Vertex AI API + ADC (for Google Cloud)
#   - auto: Use API Key if available, else try Vertex AI
NANOBANANA_AUTH_METHOD=auto

# ============================================================
# API Key Authentication (when auth_method=api_key or auto)
# ============================================================
# Recommended for local development
GEMINI_API_KEY=your_api_key_here
# GOOGLE_API_KEY=your_api_key_here  # Alternative

# ============================================================
# Vertex AI Authentication (when auth_method=vertex_ai or auto)
# ============================================================
# Google Cloud Project ID (Required for Vertex AI)
GCP_PROJECT_ID=your-project-id
# GOOGLE_CLOUD_PROJECT=your-project-id  # Alternative

# Google Cloud Region (Default: us-central1)
GCP_REGION=us-central1
# GOOGLE_CLOUD_LOCATION=us-central1  # Alternative

# ============================================================
# Model Settings
# ============================================================
# Options: flash (speed), pro (quality), auto (intelligent selection)
NANOBANANA_MODEL=auto

# ============================================================
# Image Output Directory
# ============================================================
# This is the DEFAULT directory when output_path parameter is not specified.
# The output_path parameter in generate_image() always takes precedence.
#
# Precedence (highest to lowest):
# 1. output_path parameter in generate_image() call
# 2. IMAGE_OUTPUT_DIR environment variable (below)
# 3. ~/nanobanana-images (default fallback)
#
# Default: ~/nanobanana-images
# IMAGE_OUTPUT_DIR=/path/to/output

# ============================================================
# Image Response Settings
# ============================================================
# Return full-resolution images in MCP response instead of thumbnails
# WARNING: Full images can be large (3-7MB each for 4K)
# RETURN_FULL_IMAGE=false

# Gemini 3 Pro Model Settings (optional, only applies when using Pro model)
# GEMINI_PRO_THINKING_LEVEL=high  # low, high
# GEMINI_PRO_ENABLE_GROUNDING=true  # Enable Google Search grounding