Research Paper Ingestion

You can deploy and run the Research Paper Ingestion MCP Server to autonomously discover, extract, and store insights from academic papers. It connects to arXiv and Semantic Scholar to gather papers, extracts key findings and techniques, and integrates the results into a memory system for ongoing AI learning.

How to use

You interact with this MCP server using an MCP client. The server exposes a set of actions that let you search for papers, download PDFs, extract insights, analyze citations, and store knowledge in enhanced memory. Typical usage flows include discovering relevant papers, evaluating their impact, retrieving full texts, extracting structured insights, and persisting those insights for future learning.

# Example usage pattern (conceptual - adapt to your MCP client):
# 1) Search for relevant papers
papers = mcp__research_paper__search_arxiv({
  "query": "meta-learning for few-shot learning",
  "max_results": 10
})

# 2) Download a high-impact paper by URL
paper_url = papers[0]["pdf_url"]
pdf = mcp__research_paper__download_paper({
  "url": paper_url,
  "paper_id": papers[0]["id"]
})

# 3) Extract insights from paper text or abstract
insights = mcp__research_paper__extract_insights({
  "paper_text": papers[0]["abstract"]
})

# 4) Store knowledge into enhanced memory
mcp__research_paper__store_paper_knowledge({
  "paper_metadata": papers[0],
  "insights": insights["insights"],
  "techniques": insights.get("techniques", [])
})

How to install

You need Python and pip installed on your system. Then install the required Python packages and set up the MCP server to run locally.

# 1) Navigate to the project path where the MCP server resides
cd ${AGENTIC_SYSTEM_PATH:-/opt/agentic}/agentic-system/mcp-servers/research-paper-mcp

# 2) Install dependencies
pip install -r requirements.txt

# 3) Run the MCP server (stdio config) using Python
python3 server.py

Available tools

search_arxiv

Search arXiv for papers by query, returning a list of matching papers and metadata.

search_semantic_scholar

Query Semantic Scholar to obtain metadata and citation metrics for papers.

download_paper

Download the full text PDF of a paper from a given URL and save it with a unique identifier.

extract_insights

Extract key findings, methodologies, results, and potential techniques from paper text or abstract.

analyze_citations

Analyze citation relationships and determine influence or connectivity within a citation graph.

store_paper_knowledge

Store extracted metadata, insights, and techniques into enhanced memory for long-term learning.