A walkthrough of building a deep research agent in Java using Spring AI and Browserbase, a headless browser search API powered by Exa. The author created a custom Spring Boot starter for Browserbase (since no official Java SDK exists) and built a research pipeline that generates diverse search queries via an LLM, fetches web content using Browserbase's search and fetch APIs, and synthesizes results into a structured markdown report with executive summary, key findings, themes, and open questions. The pipeline uses virtual threads for parallel fetching and supports local models via Ollama as a free alternative to OpenAI.
•16m watch time
Sort: