Deep Research AI Agent

Generate comprehensive research reports on any topic through AI-powered analysis

Deep Research is an intelligent research assistant that leveres advanced AI models to create in-depth reports on any topic. With an intuitive interface and powerful backend, it automates the research process by searching for information, analyzing content, and generating well-structured reports.

Features

• 🔍 Intelligent Topic Research: Input any research topic and get a comprehensive report
• 🤔 Interactive Clarification: Answer questions to refine the research scope
• 🔄 Real-time Progress Tracking: Monitor the research process with detailed activity logs
• 📚 Multiple AI Model Support: Choose from Gemini, OpenAI, or OpenRouter models
• 📝 Markdown Reports: Download or copy reports in clean Markdown format
• 🔗 Source Citations: All research includes source links for reference

System Architecture

flowchart LR
    subgraph Frontend
        UI[User Interface]
        Store[Zustand Store]
        UI <--> Store
    end
    
    subgraph Backend
        API[API Routes]
        RFunc[Research Functions]
        MCaller[Model Caller]
        ATracker[Activity Tracker]
        
        API --> RFunc
        RFunc --> MCaller
        RFunc --> ATracker
    end
    
    subgraph "External Services"
        EXA[Exa Search API]
        LLM[AI Models]
        subgraph "Model Providers"
            OPENAI[OpenAI]
            GOOGLE[Google Gemini]
            OPENR[OpenRouter]
        end
        
        LLM --> OPENAI
        LLM --> GOOGLE
        LLM --> OPENR
    end
    
    Store <--> API
    RFunc --> EXA
    MCaller --> LLM

Loading

Research Workflow

flowchart TB
    subgraph User
        A[User Input] --> B[Research Topic]
        B --> C{Answer Questions}
    end
    
    subgraph "AI Research Process"
        C --> D[Generate Search Queries]
        D --> E[Search for Information]
        E --> F[Extract Content]
        F --> G[Analyze Results]
        G --> H{Sufficient?}
        H -->|No| D
        H -->|Yes| I[Generate Report]
    end
    
    subgraph "Output"
        I --> J[Research Report]
        J --> K[Download/Copy]
    end
    
    C --> L[Activity Tracking]
    L --> G

Loading

Tech Stack

• Frontend: Next.js, React, Tailwind CSS, shadcn/ui components
• State Management: Zustand
• Backend: Next.js API Routes
• APIs: Exa Search API, AI model providers (OpenAI, Google, OpenRouter)
• Styling: Tailwind CSS with custom components

Getting Started

Prerequisites

• Node.js 18+
• npm or yarn
• API keys for:
  • Exa Search API
  • At least one AI model provider (OpenAI, Google Gemini, or OpenRouter)

Installation

1. Clone the repository:

git clone https://github.com/yourusername/deep-research-ai-agent.git
cd deep-research-ai-agent

2. Install dependencies:

npm install
# or
yarn install

3. Create a .env.local file in the root directory:

EXA_SEARCH_API_KEY=your_exa_api_key
OPENAI_API_KEY=your_openai_api_key
GEMINI_API_KEY=your_GEMINI_API_KEY
OPENROUTER_API_KEY=your_openrouter_api_key

4. Start the development server:

npm run dev
# or
yarn dev

5. Open http://localhost:3000 with your browser to see the application.

Configuration Options

The application allows configuration of several parameters:

• Model Provider: Choose between Gemini, OpenAI, and OpenRouter
• Max Iterations: Control the maximum research loops (default: 2)
• Max Search Results: Limit the number of search results per query (default: 1)
• Max Content Characters: Set the maximum content length from each source (default: 20000)

Usage Guide

1. Enter Research Topic: Input your research topic on the main screen
2. Answer Clarification Questions: Provide answers to help focus the research
3. Monitor Progress: Watch as the system searches for information and analyzes content
4. Review Report: Once complete, review, download, or copy the generated report

Development

Key Workflow Components

1. User Input: Collects the research topic from the user
2. Question Generation: Creates clarification questions based on the topic
3. Research Planning: Generates search queries to find relevant information
4. Content Extraction: Processes search results into structured content
5. Analysis: Determines if collected information is sufficient
6. Report Generation: Creates a comprehensive markdown report

Adding New Model Providers

To add support for a new AI model provider:

1. Update src/store/deepResearch.ts to include the new provider type
2. Add model configuration in src/app/api/deep-research/constants.ts
3. Implement the provider client in src/app/api/deep-research/services.ts
4. Update the model selection UI in src/app/page.tsx

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

• Next.js
• React
• Tailwind CSS
• shadcn/ui
• Zustand
• Exa Search API