deep_researcher
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@deep_researcherResearch the latest developments in solid-state battery technology"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Deep Researcher Agent & MCP Server ๐๐ง
An autonomous multi-stage AI research workflow agent that searches the web, analyzes unstructured data, and compiles publication-quality technical reports. Features both a Streamlit Web UI and a native Model Context Protocol (MCP) server for integration with Claude Desktop and Cursor.
Developed and maintained by Sakshi Pandey (231FA04H01@gmail.com).
๐๏ธ Multi-Stage Pipeline Workflow
graph TD
classDef inputNode fill:#1E293B,stroke:#38BDF8,stroke-width:2px,color:#F8FAFC;
classDef stageNode fill:#0F172A,stroke:#818CF8,stroke-width:2px,color:#F8FAFC;
classDef outputNode fill:#064E3B,stroke:#34D399,stroke-width:2px,color:#FFF;
Topic["๐ฏ User Research Query / Topic"]:::inputNode
subgraph AutonomousResearchPipeline ["๐ Multi-Stage Agno Workflow Engine"]
Topic --> Searcher["๐ Searcher Agent<br/><i>(Scrapegraph AI & Web Extraction)</i>"]:::stageNode
Searcher --> Analyst["๐ Analyst Agent<br/><i>(Pattern synthesis & data normalization)</i>"]:::stageNode
Analyst --> Writer["โ๏ธ Writer Agent<br/><i>(Markdown synthesis with inline citations)</i>"]:::stageNode
end
subgraph ClientInterfaces ["๐ Delivery Channels"]
Writer --> StreamlitUI["๐ป Streamlit Web UI<br/><i>(Interactive research exploration)</i>"]:::outputNode
Writer --> MCPServer["๐ Native MCP Server<br/><i>(Direct tool in Claude Desktop & Cursor)</i>"]:::outputNode
endRelated MCP server: research
๐ Key Features
๐ Multi-Stage Orchestration: Dedicated sub-agents (Searcher, Analyst, Writer) collaborate sequentially to produce exhaustive, cited research reports.
๐ AI-Assisted Scraping: Extracts data from live web pages using Scrapegraph AI and Nebius AI models.
๐ Native MCP Server: Exposes the deep research pipeline as an MCP tool directly accessible inside Claude Desktop, Cursor, or any MCP-compatible client.
๐ป Multiple Interfaces: Run research workflows via Streamlit web app, direct command line script, or through the background MCP server.
๐ ๏ธ Tech Stack
Agent Orchestrator: Agno
Inference Provider: Nebius Token Factory (Qwen / Llama models)
Web Extraction: Scrapegraph AI
Protocol: Model Context Protocol (MCP)
UI & Visualization: Streamlit
๐ Repository Structure
deep-researcher-agent-mcp/
โโโ app.py # Streamlit web interface
โโโ agents.py # Core multi-stage agent pipeline (Searcher, Analyst, Writer)
โโโ server.py # Native Model Context Protocol (MCP) server
โโโ assets/ # Architecture graphics and demo assets
โโโ pyproject.toml # uv / pip dependency specifications
โโโ .env.example # API key template
โโโ .gitignore # Git ignore rules
โโโ LICENSE # MIT Licenseโก Quick Start
1. Prerequisites
Python 3.10 or higher
uv (recommended) or
pipNebius Token Factory API Key
Scrapegraph AI API Key
2. Installation
# Clone the repository
git clone https://github.com/blue007-arc/deep-researcher-agent-mcp.git
cd deep-researcher-agent-mcp
# Install dependencies with uv
uv sync3. Environment Setup
cp .env.example .envAdd your API keys to .env:
NEBIUS_API_KEY=your_nebius_api_key_here
SGAI_API_KEY=your_scrapegraph_api_key_here๐ป Running the Agent
Option 1: Web Interface (Streamlit)
uv run streamlit run app.pyOpen http://localhost:8501 in your browser. Enter any topic (e.g. "State of Autonomous Coding Agents in 2025") and watch the multi-stage research flow stream in real-time.
Option 2: Command Line
uv run python agents.pyOption 3: Connect to Claude Desktop or Cursor (MCP)
Add the server configuration to your claude_desktop_config.json or .cursor/mcp.json:
{
"mcpServers": {
"deep_researcher": {
"command": "python",
"args": [
"run",
"server.py"
],
"env": {
"NEBIUS_API_KEY": "your_nebius_api_key_here",
"SGAI_API_KEY": "your_scrapegraph_api_key_here"
}
}
}
}๐ค Author & Maintainer
Sakshi Pandey
GitHub: @blue007-arc
Email: 231FA04H01@gmail.com
๐ License
This project is licensed under the MIT License - see the LICENSE file for details.
This server cannot be deployed
Maintenance
Related MCP Connectors
Live AI-native web search with citations. One tool for every MCP client. Flat per-request pricing.
Web research for agents: quality-scored Google search, webpage extraction, and deep research.
Scrape, crawl and search the web for AI agents via MCP.
Shared, peer-validated knowledge archive for AI agents โ search, contribute, and validate via MCP
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables deep research tasks using a multi-agent architecture that integrates any LLM and MCP tools. Available via MCP stdio, streamable HTTP, and SSE transports.17MIT
- AlicenseNot gradedqualityCmaintenanceMulti-purpose research MCP server integrating web search, deep research, web scraping, research methodology routing, and GPT Researcher report generation.MIT
- AlicenseNot gradedqualityDmaintenanceAn MCP-native document research agent that provides search, summarization, and citation tools for iterative multi-step research, ending in validated structured JSON reports.MIT
- FlicenseNot gradedqualityCmaintenanceEnables AI agents to perform unified web research through a single MCP server, including search, page fetching, recursive crawling, document parsing, YouTube transcript extraction, and deep multi-query research.3-