Open Deep Research

This podcast introduces the LangChain Deep Research Agent, an open-source tool designed to help users with complex research tasks. The agent operates in three phases: scoping the problem by clarifying user needs and creating a research brief, conducting research using a supervisor to delegate tasks to sub-agents, and writing a final report based on the accumulated findings. The agent is highly configurable, allowing users to add their own tools and customize models for different steps in the research pipeline. The podcast also provides instructions on how to run the agent locally using LangGraph Studio and introduces the Open Agent Platform as an alternative UI for trying out the Research Agent. A practical example is given, showing how the agent can plan a trip to Amsterdam and Norway, including flight options, accommodations, and a sample itinerary.