safetySettings
(array) - Controls content filtering by harm category
How to use it
safetySettings is exposed by the MCP Gemini Server MCP server. Add the server to your MCP client (Claude Desktop, Cursor, Windsurf and others), and the safetySettings tool becomes available to the model automatically. See the full listing for setup details and every tool this server provides.
FULL MCP GEMINI SERVER LISTINGOther tools in MCP Gemini Server (39)
(array) - Specific domains to allow for this request
(string enum) - Type of analysis to perform:
(string) - Identifier for cached content to use with this request
Multi-URL content comparison
(boolean) - Convert HTML to markdown (default: true)
(string) - Model to fall back to if routing fails. If not provided and routing fails, an error will be thrown.
(string) - Human-readable name for the cached content
Structured information extraction
Credibility assessment
(object) - Configuration for URL fetching
(object) - Controls generation parameters like temperature, topP, etc.
(array) - Initial conversation history
(boolean) - Include URL metadata in context (default: true)
(number) - Maximum content size per URL in KB (default: 100)
(string) - The text message to be routed to the most appropriate model
(string) - Name of the model to use
(object) - Model selection preferences
(array) - Array of model names to consider for routing (e.g., ['gemini-1.5-flash', 'gemini-1.5-pro']). The first model in the list will be used for routing decisions.
(string - features to avoid in the generated image)
(number - 1-8, default: 1)
Question-based content analysis
(string) - Simplified control: "none" (0 tokens), "low" (1K), "medium" (8K), "high" (24K)
(string enum: "512x512", "1024x1024", "1536x1536")
(string) - Custom prompt to use for routing decisions. If not provided, a default routing prompt will be used.
(number - integer value for reproducible generation)
Emotional tone analysis
(string enum: "photographic", "digital-art", "cinematic", "anime", "3d-render", "oil-painting", "watercolor", "pixel-art", "sketch", "comic-book", "neon", "fantasy")
(number - strength of style preset, 0.0-1.0)
Comprehensive content summarization
(string or object) - System instruction to guide model behavior
(number) - Maximum tokens for reasoning (0-24576)
(object) - Controls model reasoning process
(number) - Fetch timeout per URL in milliseconds (default: 10000)
(object) - Configures tool behavior like temperature and confidence thresholds
(array) - Tool definitions including function declarations
(string - e.g., '3600s') - Time-to-live for the cached content
(object) - Fetch and include web content from URLs
(array) - URLs to fetch and include as context (max 20)
(string) - Custom User-Agent header for URL requests