🔍 Autonomous Deep Research System

An AI-powered research assistant built with Python, OpenAI Agents SDK, and Gradio that autonomously plans research, performs parallel web searches, synthesizes findings into a structured report, and emails the final results.

---

Overview

This project demonstrates an autonomous multi-stage research workflow where specialized AI agents collaborate to answer complex research questions.

Instead of relying on a single LLM response, the system decomposes the problem into multiple research tasks, executes searches concurrently, generates a comprehensive report, and automatically delivers the results via email.

---

Demo

---

Docker

Build

docker build -t deep-research-agent .

Run

docker run -p 7860:7860 ^
-e OPENAI_API_KEY=YOUR_OPENAI_API_KEY ^
deep-research-agent

Open

http://localhost:7860

---

Features

• Autonomous research planning
• Parallel web search execution
• AI-powered report generation
• Email delivery of research reports
• Interactive Gradio web interface
• OpenAI tracing for workflow observability
• Modular agent architecture

---

Architecture

User Question
      │
      ▼
Planner Agent
      │
      ▼
Search Plan
      │
      ▼
Multiple Search Agents
      │
      ▼
Collected Results
      │
      ▼
Writer Agent
      │
      ▼
Research Report
      │
      ▼
Email Agent
      │
      ▼
Delivered Report

---

Technology Stack

• Python
• OpenAI Agents SDK
• OpenAI API
• Gradio
• Pydantic
• Python Dotenv
• Requests

---

Installation

Clone the repository

git clone https://github.com/NeelPawar-01/deep-research-agent.git

Move into the project

cd deep-research-agent

Install dependencies

pip install -r requirements.txt

Create a .env file

OPENAI_API_KEY=your_api_key

Run the application

python app.py

---

Example Workflow

1. Enter a research question.
2. The Planner Agent creates a research strategy.
3. Search Agents perform parallel information gathering.
4. The Writer Agent synthesizes the findings into a structured report.
5. The Email Agent sends the completed report.
6. The user receives both the report in the UI and via email.

---

Skills Demonstrated

• Autonomous AI Agents
• Agent Orchestration
• Parallel Task Execution
• OpenAI Agents SDK
• LLM Workflow Design
• Research Automation
• Python Application Development
• Async Programming
• Prompt Engineering

---

Future Improvements

• Support for local LLMs
• Retrieval-Augmented Generation (RAG)
• PDF export
• Citation management
• Multi-provider LLM support
• Persistent research history

---

About

This project demonstrates how autonomous AI agents can collaborate to perform complex research tasks through planning, execution, synthesis, and automated report delivery.