Intelligent Web & Local Search Agent with Brave API & Gemini MCP

This n8n workflow creates an AI-powered search agent that processes natural language queries using Google Gemini (gemini-2.5-flash-preview-05-20). It intelligently selects between Brave's web search and local search tools via external MCP services (e.g., Smithery), orchestrating them seamlessly as a single callable MCP tool named 'call_brave_search_agent'. The modular design includes an MCP server trigger, AI agent node, Gemini chat model, workflow tools, and MCP client tools for integration.

Key benefits include simplified integration into other apps or agent systems, reduced client-side LLM token costs by offloading logic to n8n, centralized management of search strategies, and extensibility for additional tools. It supports conversational memory within executions and customizable system prompts for tailored behavior.

Ideal use cases: Enhancing AI agents with advanced search in chatbots, automating research tasks, building RAG systems, or integrating into development tools for IT teams. Perfect for developers needing robust, AI-driven search without complex client-side implementations.