🧠 DeepSearch – AI Research Agent

DeepSearch is an AI-powered research assistant that helps users generate structured research reports by first asking intelligent clarifying questions and then producing a comprehensive analysis.

Unlike simple chatbots, DeepSearch behaves like an AI research agent, planning and generating research based on user intent.

---

🚀 Features

• 🤖 AI-generated clarifying questions
• 🧩 Multi-step research pipeline
• 📊 Live research activity tracking
• ⏱️ Research time indicator
• 🌗 Light / Dark theme (desktop & mobile)
• 📄 Structured research report output
• ⚡ Real-time streaming responses
• 🧠 Free-mode support (no external search required)

---

🛠 Tech Stack

• Next.js 14 (App Router)
• TypeScript
• AI SDK (ai)
• OpenRouter (LLaMA models)
• Tailwind CSS
• ShadCN UI
• Zod
• Vercel

---

📂 Project Structure

src/
├── app/
│ ├── api/
│ │ ├── generate-questions/
│ │ └── deep-research/
│ ├── layout.tsx
│ └── page.tsx
├── components/
│ ├── ui/
│ │ └── deep-research/
│ ├── ThemeProvider.tsx
│ └── ThemeToggle.tsx
├── store/
│ └── deepResearch.ts

---

⚙️ Setup & Run Locally

Clone the repo

git clone https://github.com/AadiSharma49/DeepSearch-AI-Agent.git
cd DeepSearch-AI-Agent

Install dependencies

npm install

Add environment variables

Create .env.local

OPENROUTER_API_KEY=your_api_key_here

Run dev server

npm run dev

Build for production

npm run build

---

Deployment

This project is Vercel-ready.

---

Future Improvements

• Cache questions per topic
• Regenerate questions button
• User-selected question count
• Authentication (login)
• Save research history
• Server-side PDF export
• Faster (<500ms) question generation

---

Author

Aaditya Sharma