An intelligent multi-agent research system that transforms web research through AI orchestration
Revolutionizing research workflows with autonomous AI agents
Deep Research AI Agent is a sophisticated research automation system that employs a multi-stage AI workflow to conduct comprehensive web research. Built with cutting-edge AI frameworks, it orchestrates three specialized agents to deliver publication-quality research reports in minutes, not hours.
- 🤖 Multi-Agent Architecture: Three specialized AI agents working in perfect harmony
- 🔍 Intelligent Web Scraping: Advanced extraction using ScrapeGraph AI technology
- 📊 Real-time Analysis: Live synthesis and interpretation of research findings
- 📝 Automated Report Generation: Publication-ready reports with proper citations
- 🌐 Streamlit Interface: Intuitive web application for seamless user experience
- 🔌 MCP Integration: Model Context Protocol server for enhanced AI tool integration
The system employs a three-stage pipeline powered by Nebius AI Studio and ScrapeGraph AI:
graph LR
A[🔍 Searcher Agent] --> B[🔬 Analyst Agent] --> C[✍️ Writer Agent]
A --> D[Web Sources]
B --> E[Synthesized Insights]
C --> F[📄 Final Report]
| Agent | Responsibility | Technology Stack |
|---|---|---|
| 🔍 Searcher | Discovers and extracts web information | ScrapeGraph AI + Qwen-30B |
| 🔬 Analyst | Synthesizes findings and identifies insights | Qwen-30B + Critical Analysis |
| ✍️ Writer | Produces publication-quality reports | Qwen-14B + Technical Writing |
- Autonomous Research: End-to-end research automation without human intervention
- Multi-Source Intelligence: Aggregates data from news, blogs, papers, and forums
- Real-time Processing: Live status updates and streaming report generation
- Citation Management: Automatic source verification and reference formatting
- Controversy Detection: Identifies conflicting viewpoints and debates
- Scalable Architecture: Modular design supporting easy agent customization
- Error Resilience: Robust error handling and graceful degradation
- Streaming Interface: Real-time content delivery for enhanced user experience
- API Integration: RESTful endpoints and MCP server compatibility
- Security First: API key encryption and secure credential management
- Python 3.10+
- Nebius AI Studio API key
- ScrapeGraph AI API key
# Clone the repository
git clone https://github.com/priyankeshh/deep-research-agent.git
cd deep-research-agent
# Install dependencies using uv (recommended)
uv sync
# Or use pip
pip install -r requirements.txtCreate a .env file in the project root:
NEBIUS_API_KEY=your_nebius_api_key_here
SGAI_API_KEY=your_scrapegraph_api_key_here# Using uv
uv run streamlit run app.py
# Or using Python directly
streamlit run app.pyNavigate to http://localhost:8501 to access the research interface.
# Technology Research
"Analyze the current state of quantum computing and its commercial applications"
# Market Analysis
"Compare the adoption rates of electric vehicles across different regions"
# Trend Investigation
"Investigate the impact of AI on software development productivity"from agents import DeepResearcherAgent
# Initialize the research agent
agent = DeepResearcherAgent()
# Execute research workflow
research_results = agent.searcher.run("Your research query")
analysis = agent.analyst.run(research_results.content)
report = agent.writer.run(analysis.content)
print(report.content)# Start MCP server
python server.py
# Add to your .cursor/mcp.json
{
"mcpServers": {
"deep_researcher_agent": {
"command": "uv",
"args": ["run", "python", "server.py"],
"env": {
"NEBIUS_API_KEY": "your_key_here",
"SGAI_API_KEY": "your_key_here"
}
}
}
}- Agno: Advanced AI agent orchestration framework
- ScrapeGraph AI: Intelligent web scraping and data extraction
- Nebius AI Studio: High-performance language model inference
- Streamlit: Interactive web application framework
- Model Context Protocol (MCP): AI tool integration standard
- Pydantic: Data validation and settings management
- Python-dotenv: Environment variable management
- Multi-Agent Orchestration: Sophisticated workflow coordination between specialized AI agents
- Real-time Streaming: Live report generation with progressive content delivery
- Source Verification: Automated fact-checking and citation validation
- Modular Architecture: Easily extensible system supporting custom agent implementations
- Zero-Configuration Setup: Streamlined installation and deployment process
- Enterprise-Ready: Scalable architecture supporting high-throughput research operations
- API-First Design: RESTful endpoints enabling seamless integration
- Cross-Platform Compatibility: Consistent performance across different operating systems
Each agent can be customized through environment variables and configuration files:
# Custom model configuration
searcher = Agent(
model=Nebius(id="custom-model", api_key=api_key),
tools=[ScrapeGraphTools(api_key=sgai_key)],
description="Custom searcher instructions..."
)- Model Selection: Choose optimal models for each agent based on task complexity
- Concurrency Control: Adjust parallel processing limits for your infrastructure
- Cache Management: Configure result caching for improved response times
Contributions are welcome! This project follows modern development practices:
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
# Install development dependencies
uv add --dev pytest black flake8 mypy
# Run tests
uv run pytest
# Format code
uv run black .
# Type checking
uv run mypy .This project is licensed under the MIT License - see the LICENSE file for details.
Priyankesh
- 📧 Email: priyankeshom@gmail.com
- 🐙 GitHub: @priyankeshh
- 🔗 LinkedIn: Priyankesh
- Agno Team for the excellent AI agent framework
- ScrapeGraph AI for advanced web scraping capabilities
- Nebius for high-performance AI model inference
- Streamlit for the intuitive web application framework
⭐ Star this repository if you find it useful!
