AI Web Search API with Brave Search & Gemini Semantic Re-Ranking

This n8n workflow builds a callable web-research API that goes beyond a plain search engine. Incoming queries are first optimized by Google Gemini, then sent to Brave Search to gather raw results, which are semantically re-ranked by an LLM (with Anthropic Claude and OpenAI available as alternative models) so the most relevant sources rise to the top before the final answer is returned via webhook. It's designed as a reusable research microservice that other workflows, agents, or apps can call whenever they need accurate, ranked web search results rather than a raw list of links.

Category: Data Extraction & Web Scraping