🤖 LLM-Agent-ResearchSuite

An AI-powered research assistant built using LangChain, Gemini 1.5/2.5, GPT-4, Claude 3.5, and real-time web tools like DuckDuckGo and Wikipedia.

This project enables structured research generation with support for saving, summarizing, and exporting insights in a clean, user-friendly format.

---

🧩 Problem Statement

With the explosion of online content, manually researching any topic is time-consuming. This agent automates the research process using LLMs and tools to generate structured summaries with verified sources — all in seconds.

---

🚀 Features

• 🔍 Multi-tool Research Agent: Combines Wikipedia, DuckDuckGo search, and file-saving capabilities using LangChain Agents.
• 🧠 LLM Integration: Supports top-tier models (Gemini 1.5/2.5, GPT-4, Claude 3.5) for accurate and structured output.
• 🧾 Structured Results: Automatically returns research in a standardized JSON format (Topic, Exploration, Summary, Sources, Tools Used).
• 📄 PDF Export: Download clean PDF reports generated from markdown using xhtml2pdf.
• 🖱️ No-code UI: Built with Streamlit for fast, interactive use.

---

🌐 Live Demo

Check out the live version here:
👉 LLM-Agent-ResearchSuite – Streamlit App

⚠️ Note: This app may take a few seconds to load based on Streamlit’s server spin-up time.

---

📦 Tech Stack

Component	Description
LangChain	Agent & tool orchestration
Gemini/GPT/Claude	Large Language Models for research
DuckDuckGo API	Real-time web search tool
Wikipedia API	Fast lookup for verified facts
Streamlit	Front-end interface
Pydantic	Structured output validation
xhtml2pdf	PDF generation from markdown

---

🧠 How It Works

1. User enters a research topic.
2. LLM Agent is activated with selected tools (e.g., Wikipedia, Web Search).
3. LangChain AgentExecutor loops through tools to gather relevant info.
4. Output is structured, summarized, and can be saved/exported.

---

📁 Folder Structure

.
├── tools.py          # LangChain tool definitions
├── web.py            # Main Streamlit app + agent pipeline
├── README.md
└── requirements.txt  # Dependencies

---

📸 Screenshots

Here are a few screenshots that demonstrate how the app works:

🏠 Home Screen

Shows the initial UI where the user enters a research query, selects tools, and picks an LLM.

---

📋 Research Output

Displays the structured research result, including topic, exploration, summary points, sources, and tools used.

---

📄 Downloaded PDF Preview

Highlights the PDF generation feature, allowing users to download a clean, formatted research report.

---

🧪 Example Use Cases

• 📚 Academic Research Summaries
• 📈 Market or Business Analysis
• 🦸 Fun Topics (e.g., "Unknown Facts about Batman")
• 🔍 Real-time info gathering + PDF reporting

---

🎯 Target Audience

• Students writing reports or research essays
• Content creators looking to outline factual content
• Developers learning how to integrate LLMs with tools

---

✅ Why This Project Stands Out

• ✔️ Combines multi-agent orchestration and tool calling — not just a chatbot
• ✔️ Focused on structured knowledge, not random chat
• ✔️ Versatile for both fun and professional use
• ✔️ Strong backend + frontend integration

---

🚫 Known Limitations

• Heavily depends on the quality and availability of search results.
• Requires API keys; free-tier models may sometimes return incomplete outputs.
• Currently supports only English input and output.

---

🧪 Evaluation

The agent was tested across 10 diverse topics (tech, history, science). It consistently returned structured summaries with 85–90% relevance and accuracy when manually compared to top search results.

---

🙋‍♂️ About Me

I’m a passionate Computer Science student and aspiring AI engineer. This project showcases my ability to combine:

• 🤖 AI agent design
• 🧰 Tool integration (LangChain ecosystem)
• 🖥️ Full-stack development with Python + Streamlit
• 📚 Prompt engineering & structured output parsing

---

📎 How to Run Locally

1. Clone the repo

git clone https://github.com/Yaser-123/LLM-Agent-ResearchSuite.git
cd LLM-Agent-ResearchSuite

2. Install dependencies

pip install -r requirements.txt

3. Add your .env file with your keys:

GOOGLE_API_KEY=your_key
OPENAI_API_KEY=your_key
ANTHROPIC_API_KEY=your_key

4. Run the app

streamlit run web.py

---

💡 Future Improvements

• Add tool usage memory (LangGraph or ReAct-style)
• Add more search engines (Google SERP, You.com)
• Allow citation formatting (APA/MLA/Harvard)
• Cloud storage support for research archives

---

⭐ Give It a Star

If you found this project useful or inspiring, consider giving it a ⭐️ on GitHub!

---

📫 Contact

Feel free to connect with me on LinkedIn or reach out via email at [email protected] for collaboration opportunities.