Deep Research Chatbot using OpenAI

This repository contains a Jupyter notebook demonstrating a fully automated "Deep Research" agent built with the OpenAI Agents SDK. The agent can plan web searches, summarize findings, compose a cohesive report, and email the results—all with minimal user intervention.

Features

• Automated Planning & Search: Uses a Planner agent to break down research queries into targeted web searches and a Search agent to retrieve concise summaries.
• Report Generation: Leverages a Writer agent to synthesize search summaries into a structured, 2-3 paragraph report under 300 words.
• Email Delivery: Implements an Email agent powered by SendGrid to automatically send the generated report to a specified recipient.
• Traceability: Integrates OpenAI’s trace utility for end-to-end visibility into each step of the agent’s decision-making.

Notebook Structure

1. Setup & Imports: Load environment variables, import SDK components, and configure API clients.

2. Agent Definitions:

   • search_agent: Summarizes web search results.
   • planner_agent: Plans a series of targeted searches based on the user’s query.
   • writer_agent: Drafts the final research report.
   • email_agent: Composes and sends the report via SendGrid.

3. Planning & Execution Functions:

   • plan_searches(): Generates a search plan.
   • perform_searches(): Executes searches and collects summaries.

4. Report & Email Functions:

   • write_report(): Generates the report from search summaries.
   • send_email(): Sends the report as an email.

5. Execution Example: Runs a full research cycle on a sample query, e.g., "Latest AI Agent frameworks in 2025."

6. Tracing: View detailed run logs on the OpenAI Traces dashboard.

Prerequisites

• Python 3.8+
• An OpenAI API key with Agents SDK access
• A SendGrid API key for email delivery

Installation

1. Clone the repository:

   git clone https://github.com/Rishi-Kora/Deep-Research-chatbot-using-OpenAI.git
   cd Deep-Research-chatbot-using-OpenAI

2. Install dependencies:

   pip install agents-sdk python-dotenv pydantic sendgrid

3. Copy .env.example to .env and fill in your API keys:

   OPENAI_API_KEY=your_openai_key
   SENDGRID_API_KEY=your_sendgrid_key
   EMAIL_SENDER=your_verified_sendgrid_sender@example.com
   EMAIL_RECIPIENT=recipient@example.com

Usage

1. Open the Jupyter notebook:

   jupyter notebook Deep_research_using_OpenAI.ipynb

2. Update the query variable in the final cell with your own research topic.

3. Run all cells. The notebook will:

   • Plan and execute web searches
   • Generate a concise research report
   • Email the report

License

This project is licensed under the MIT License.