Brave Search MCP

Large language models are powerful reasoners, but their knowledge has a hard cutoff date. Brave Search MCP solves this by letting your AI agent query the open web in real time through Brave's independent search index. When your agent needs to verify a fact, look up recent documentation, or find current pricing, it can search the web directly and incorporate fresh results into its response, no browser required. The server exposes two focused tools: web search for general queries and local search for location-based results. This simplicity is intentional. Rather than wrapping an entire browser automation framework, Brave Search MCP does one thing well: it retrieves relevant, up-to-date information from the web and delivers it in a structured format your agent can reason over. The Brave Search API also operates on an independent index, not resold Google results, which means you sometimes get different and complementary perspectives on a topic. Getting started takes under five minutes. Sign up for a free Brave Search API key, set the environment variable, and add the server to your MCP client. The free tier is generous enough for personal development use, and paid plans scale for heavier workloads. Because the server runs locally via stdio, your queries never pass through a third-party proxy beyond Brave's own API.