Skip to main content
Run a Paper Search agent and chatbot at the Room level in the Meshagent cloud for easy testing. 1. Install Meshagent CLI 
pip install "meshagent[all]"
2. Sign up and Authenticate 3. Start MCP Paper Search Server in a Room 
meshagent service test --room=test --role=agent --image=meshagent/mcp-paper-search:latest --env MESHAGENT_PORT=8001 --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" --name=mcp-paper-search-service-test
  • This launches a Meshagent Room with the Paper Search MCP server available. Inactive Rooms (and services) are cleaned up automatically.
4. Start a Chatbot in the Room Using the Toolkit 
meshagent chatbot join --room=test --agent-name=mcp-paper-search --name=mcp-paper-search --toolkit=mcp-paper-search
  • This creates a chatbot in the Room bound to the MCP server. Use multiple toolkits per agent if desired.
  • A Room link appears in the output.
5. Try it Out!
  • Open the Room link in your browser and message the agent to interact with the MCP server tools in real-time.

​
Project-Level Deployment (Persistent Services)

For production: create persistent MCP Server and chatbot services at the project level. These join every new Room in your Meshagent project automatically—no need for manual process management. 1. Create Persistent MCP Server 
meshagent service create --role=agent --image=meshagent/mcp-paper-search:latest --env MESHAGENT_PORT=8001 --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" --name=mcp-paper-search-service
2. Create Persistent Chatbot 
meshagent service create --image="meshagent/cli:latest" --port="num=9001 path=/agent liveness=/ type=meshagent.callable participant_name=mcp-paper-search-chatbot" --env="MESHAGENT_PORT=9001" --name="mcp-paper-search-chatbot-service" --command="meshagent chatbot service --agent-name=mcp-paper-search-chatbot --toolkit=mcp-paper-search"
  • Once deployed, these will show up in all Meshagent project Rooms upon creation.

​
Tools Available

Paper Search MCP Server provides academic paper search and download functionality from sources like arXiv, PubMed, bioRxiv, and more.
ToolDescription
download_arxivDownload PDF of an arXiv paper.
download_biorxivDownload PDF of a bioRxiv paper.
download_medrxivDownload PDF of a medRxiv paper.
download_pubmedAttempt to download PDF of a PubMed paper.
read_arxiv_paperRead and extract text content from an arXiv paper PDF.
read_biorxiv_paperRead and extract text content from a bioRxiv paper PDF.
read_medrxiv_paperRead and extract text content from a medRxiv paper PDF.
read_pubmed_paperRead and extract text content from a PubMed paper.
search_arxivSearch academic papers from arXiv.
search_biorxivSearch academic papers from bioRxiv.
search_google_scholarSearch academic papers from Google Scholar.
search_medrxivSearch academic papers from medRxiv.
search_pubmedSearch academic papers from PubMed.

​
Tools Details

​
download_arxiv

Description: Download PDF of an arXiv paper.
Parameters:
  • paper_id (string): arXiv paper ID (e.g., ‘2106.12345’).
  • save_path (string, optional): Directory to save the PDF (default: ./downloads).

​
download_biorxiv

Description: Download PDF of a bioRxiv paper.
Parameters:
  • paper_id (string): bioRxiv DOI.
  • save_path (string, optional): Directory to save the PDF (default: ./downloads).

​
download_medrxiv

Description: Download PDF of a medRxiv paper.
Parameters:
  • paper_id (string): medRxiv DOI.
  • save_path (string, optional): Directory to save the PDF (default: ./downloads).

​
download_pubmed

Description: Attempt to download PDF of a PubMed paper.
Parameters:
  • paper_id (string): PubMed ID (PMID).
  • save_path (string, optional): Directory to save the PDF (default: ./downloads).

​
read_arxiv_paper

Description: Read and extract text content from an arXiv paper PDF.
Parameters:
  • paper_id (string): arXiv paper ID (e.g., ‘2106.12345’).
  • save_path (string, optional): Directory where the PDF is/will be saved (default: ./downloads).

​
read_biorxiv_paper

Description: Read and extract text content from a bioRxiv paper PDF.
Parameters:
  • paper_id (string): bioRxiv DOI.
  • save_path (string, optional): Directory where the PDF is/will be saved (default: ./downloads).

​
read_medrxiv_paper

Description: Read and extract text content from a medRxiv paper PDF.
Parameters:
  • paper_id (string): medRxiv DOI.
  • save_path (string, optional): Directory where the PDF is/will be saved (default: ./downloads).

​
read_pubmed_paper

Description: Read and extract text content from a PubMed paper.
Parameters:
  • paper_id (string): PubMed ID (PMID).
  • save_path (string, optional): Directory where the PDF would be saved (unused).

​
search_arxiv

Description: Search academic papers from arXiv.
Parameters:
  • query (string): Search query string (e.g., ‘machine learning’).
  • max_results (integer, optional): Maximum number of papers to return (default: 10).

​
search_biorxiv

Description: Search academic papers from bioRxiv.
Parameters:
  • query (string): Search query string (e.g., ‘machine learning’).
  • max_results (integer, optional): Maximum number of papers to return (default: 10).

​
search_google_scholar

Description: Search academic papers from Google Scholar.
Parameters:
  • query (string): Search query string (e.g., ‘machine learning’).
  • max_results (integer, optional): Maximum number of papers to return (default: 10).

​
search_medrxiv

Description: Search academic papers from medRxiv.
Parameters:
  • query (string): Search query string (e.g., ‘machine learning’).
  • max_results (integer, optional): Maximum number of papers to return (default: 10).

​
search_pubmed

Description: Search academic papers from PubMed.
Parameters:
  • query (string): Search query string (e.g., ‘machine learning’).
  • max_results (integer, optional): Maximum number of papers to return (default: 10).
⌘I