Player not loading? Watch on YouTube
Uploaded July 16, 2025, this tutorial explains Open Deep Research and its local development setup. The source names no software version, so the setup reflects the upload period rather than a current installation guide.
The open source AI agent uses LangGraph to scope a request, conduct research, and write a report. It can ask clarifying questions before condensing the conversation into a research brief. A supervisor then delegates suitable subtopics to separate researchers or keeps the work in one thread. Each researcher calls tools and cleans its findings before returning them to the supervisor, which can request more research.
The speaker says this cleaning step reduces irrelevant tool output and context pressure. They also report that generating the final report in one request produced more cohesive results than writing sections in parallel. A vacation-planning example illustrates the workflow with flight searches, accommodation options, booking links, and cited sources.
The local setup walkthrough covers cloning the repository, creating a virtual environment, installing dependencies, and supplying API keys before starting a development server through LangGraph Studio. The demonstrated defaults use OpenAI models and a search service, so running the agent locally still involves external APIs. Configuration allows different models for research and writing, plus MCP tools. Open Agent Platform provides another interface for trying the agent with API keys.
