Building Deep Research Agents with the OpenAI Agents SDK

Last updated on Feb 2, 2026

This three-part tutorial series will walk you through building a durable, multi-agent deep research application with human-in-the-loop capabilities using Temporal and the OpenAI Agents SDK.

• Part 1: Setting the Stage: Clone the template repository (a non-durable research agent), run it, and understand how the multi-agent pipeline works.

• Part 2: Creating the Workflow: Build the research manager to orchestrate agents and the workflow to manage state and human-in-the-loop interactions.

• Part 3: Running Your Application: Create the Temporal Worker with the OpenAI Agents plugin, and test durability by surviving crashes.

Sign up here to get notified when we drop new educational content!

📄️ Part 1: Setting the Stage

Learn how to build a durable, multi-agent research application with human-in-the-loop capabilities using Temporal

📄️ Part 2: Creating the Workflow

Create the Temporal Workflow, Worker, and server for your durable deep research agent

📄️ Part 3: Running Your Deep Agent

Create the Temporal Workflow, Worker, and server for your durable deep research agent