# Deep Research AI Agent Generate comprehensive research reports on any topic in seconds. ## What is This? Deep Research AI Agent is a web application that helps you research any topic and generates professional research reports automatically. Simply type in what you want to research, and our AI agents will: - Search the web for reliable information - Analyze and synthesize the data - Create a structured research report - Let you download it in multiple formats (PDF, Word, Markdown) **No technical knowledge required!** Just visit the website and start researching. ## Live Demo [https://deep-research-ai-agent.streamlit.app/](https://deep-research-ai-agent.streamlit.app/) ## See It In Action [](https://vimeo.com/1076886152) ## Key Features ### Intelligent Research System - **Dual AI Agents**: One searches the web, another writes your report - **Multi-Language Support**: Generate reports in English, Spanish, or German - **Model Selection**: Choose models via OpenRouter ### Customizable Output - **Writing Styles**: Academic, Business, Technical, or Casual - **Citation Formats**: APA, MLA, or IEEE standards - **Word Count Control**: 500 to 5000 words - **Multiple Export Formats**: PDF, Word, Markdown, JSON, or Plain Text ### User-Friendly Interface - **No Login Required**: Start researching immediately - **Progress Tracking**: Real-time updates as your research generates - **Mobile Responsive**: Works on phones, tablets, and desktops ### Advanced Features - **Deep Research Mode**: For comprehensive, academic-style papers - **Model Selection**: Choose from multiple models - **Duplicate Detection**: Automatically removes redundant content - **Memory System (ChromaDB)**: Reduces API calls by 30-60%, speeds up responses by 20-40% on repeated topics ## Architecture (agent‑like) - Orchestrated with LangGraph as a two‑node state machine: - \`research\` → gathers sources via Tavily (with domain filtering) - \`draft\` → composes structured Markdown based on style/language/citations - Post‑processing normalizes lists, paragraph spacing, and references across PDF/Word/Markdown/Text. - Optional vector memory stores past research with smart TTL (3 days for news, 30 days for evergreen content) ## Perfect For - **Students**: Research papers, essays, assignments - **Writers**: Article research, fact-checking, content ideas - **Educators**: Lesson planning, curriculum development - **Anyone Curious**: Learn about any topic quickly! ## For Developers ### Quick Setup 1. **Clone the repository:** \`\`\`bash git clone https://github.com/saksham-jain177/AI-Agent-based-Deep-Research.git cd AI-Agent-based-Deep-Research \`\`\` 2. **Install Python 3.8+ and dependencies:** \`\`\`bash pip install -r requirements.txt \`\`\` 3. **Get API keys:** - [Tavily API](https://tavily.com) - For web search (free tier available) - [OpenRouter API](https://openrouter.ai) - For AI models 4. **Create \`.env\` file:** \`\`\` TAVILY_API_KEY=your_tavily_key_here OPENROUTER_API_KEY=your_openrouter_key_here # Optional: Enable vector memory for faster repeated searches ENABLE_VECTOR_STORE=true # Optional: For feedback system (bot email sends to itself) FEEDBACK_BOT_EMAIL=your_bot@gmail.com # Bot Gmail account FEEDBACK_BOT_PASSWORD=16_char_app_password # Gmail App Password (not regular password) \`\`\` 5. **Run the app:** \`\`\`bash streamlit run app.py \`\`\` ### Tech Stack - **Frontend**: Streamlit (Python web framework) - **AI Framework**: LangChain & LangGraph - **Web Search**: Tavily API - **LLM Provider**: OpenRouter - **Document Generation**: ReportLab (PDF), python-docx (Word) ### Project Structure \`\`\` AI-Agent-based-Deep-Research/ ├── app.py # Main Streamlit application ├── main.py # Orchestrates research workflow ├── research_agent.py # Web search functionality ├── draft_agent.py # AI report generation ├── requirements.txt # Python dependencies └── .env # API keys (create this) \`\`\` ## Deploy your own instance If you'd like to deploy your own version of this app with customizations with Streamlit Cloud : 1. Fork this repository 2. Visit [share.streamlit.io](https://share.streamlit.io) 3. Connect your GitHub account 4. Deploy your forked repo 5. Add API keys in Streamlit's Secrets (Settings → Secrets) ## Contributing We welcome contributions! Here's how you can help: 1. **Report Bugs**: Open an issue describing the problem 2. **Suggest Features**: Share your ideas in discussions 3. **Submit Code**: Fork, modify, and create a pull request 4. **Improve Docs**: Help make this README even better 5. **Share**: Tell others about this project! ### Development Setup \`\`\`bash # Clone your fork git clone https://github.com/YOUR_USERNAME/AI-Agent-based-Deep-Research.git # Create virtual environment python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate # Install in development mode pip install -r requirements.txt # Make your changes and test streamlit run app.py \`\`\` ## FAQ **Q: Do I need coding knowledge?**\ A: No. Open the web app, enter a query, and click Run. **Q: Can I use this for academic work?** \ A: Yes, but always verify sources and cite appropriately. This is a research tool, not a substitute for critical thinking. **Q: How accurate is the information?**\ A: We search reputable sources and filter out social media. However, always fact-check important information. **Q: Can I customize the AI model?**\ A: Yes. Choose a model from the sidebar (OpenRouter). **Q: Is my data private?**\ A: We don't store your searches. API providers may have their own policies. ## License MIT License - Use freely for personal or commercial projects! ## Contact - **GitHub Issues**: [Report bugs or request features](https://github.com/saksham-jain177/AI-Agent-based-Deep-Research/issues) - **Discussions**: [Join the community](https://github.com/saksham-jain177/AI-Agent-based-Deep-Research/discussions) ---