A modular, end-to-end agent-based research automation system powered by LLMs. This assistant autonomously plans web searches, summarizes findings, writes detailed reports, and emails the result — all from a single user query.
- 🗺️ Planner Agent: Breaks the query into structured web searches with reasoning.
- 🔍 Search Agent: Performs web searches and summarizes results concisely.
- 📝 Writer Agent: Synthesizes findings into a well-formatted markdown report.
- 📧 Email Agent: Converts reports to HTML and sends via SendGrid.
- 🌐 Gradio UI: Clean interface to enter queries and track research progress.
- 🧵 Trace Logging: Generates OpenAI trace links for debugging and transparency.
git clone https://github.com/AnishNehete/Deep-Research-using-Open-AI/
- Install dependencies
pip install -r requirements.txt
- Configure environment variables Create a .env file in the root directory:
OPENAI_API_KEY=your_openai_key SENDGRID_API_KEY=your_sendgrid_key
- Launch the app
python deep_research.py 🧠 Agent Architecture Agent Name Purpose PlannerAgent Generates 5 web search terms based on the input query SearchAgent Performs search + summarization for each generated term WriterAgent Creates a detailed markdown report with outline and summary EmailAgent Formats the report in HTML and sends it via email
📂 Code Structure File Description deep_research.py Main UI and app launcher using Gradio research_manager.py Controls the full agentic workflow planner_agent.py Query decomposition into web search plan search_agent.py Executes searches and extracts summaries writer_agent.py Converts search findings into a full report email_agent.py Sends reports via SendGrid email service
💡 Example Query
What are the ethical concerns and regulatory frameworks around AI in healthcare? You’ll receive a ~1000 word markdown report with:
A short summary
Full report content
Follow-up research questions
Delivered directly via email
🔐 Security Note All API keys must be stored in the .env file. Do not commit your credentials.
🔧 Future Enhancements Allow user to specify recipient email
Add PDF export support
Save report history to a local or cloud database
👨💻 Author Built by Anish Nehete Inspired by Ed Donner Check out his Agentic AI course :)