This paper introduces an agentic system that transforms how researchers conduct scientific literature reviews by leveraging LLMs with a patient, iterative approach.

• System independently searches for relevant literature and distills key information
• Overcomes LLM limitations through patient, multi-step processing of complex scientific questions
• Demonstrates strong performance on biology-related queries that require domain expertise
• Enables more efficient scientific research by automating literature synthesis and analysis

For biology researchers, this technology promises to accelerate discovery by efficiently navigating vast scientific literature, extracting domain-specific insights, and identifying connections between findings that might otherwise remain undiscovered.

Patience is all you need! An agentic system for performing scientific literature review