MedAdapt Content

You deploy the MedAdapt Content Server to connect Claude Desktop with medical knowledge sources, enabling fast searching, retrieval, and analysis of educational resources from PubMed, NCBI Bookshelf, and user documents to support AI-assisted medical learning.

How to use

You interact with the MedAdapt MCP through your Claude Desktop setup. Start the server locally and configure Claude to talk to it using the provided MCP endpoint. Once connected, you can search for medical content across PubMed and Bookshelf, fetch full articles or book chapters, generate topic overviews, and build structured learning plans based on your current level.

How to install

Prerequisites: Python, a functioning virtual environment, and network access to install dependencies.

# Clone the server repository
git clone https://github.com/ryoureddy/medadapt-content-server.git
cd medadapt-content-server

# Create and activate a virtual environment (recommended)
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# (Optional) Get an NCBI API key for higher rate limits and create a .env file from the example

Run the server locally with Python to start listening for MCP requests.

python content_server.py

Configure Claude Desktop to connect to the MCP server by editing the MCP configuration JSON in your environment. Use the following structure to define the mediation endpoint and runtime details.

{
  "mcpServers": {
    "medadapt": {
      "command": "/path/to/python",
      "args": [
        "/path/to/medadapt-content-server/content_server.py"
      ],
      "env": {
        "DB_PATH": "/path/to/medadapt-content-server/medadapt_content.db"
      }
    }
  }
}

Additional configuration and troubleshooting

Populate initial topic mappings to enable organized topic-specific results.

python populate_topics.py

If you need to verify functionality or catch issues early, run tests provided with the project.

python test_server.py

Common issues and fixes include ensuring the DB_PATH is an absolute path with write access, using an API key for PubMed/Bookshelf if rate limits occur, and confirming the MCP server is running and properly configured in Claude Desktop.

Available tools

search_medical_content

Search for medical content with filters across PubMed and NCBI Bookshelf to locate relevant articles, chapters, and documents.

get_resource_content

Retrieve the complete content for a specific resource, such as an article or book chapter, for in-depth study.

get_topic_overview

Generate a comprehensive overview of a medical topic, summarizing key concepts, pathways, and current knowledge.

suggest_learning_resources

Provide personalized resource recommendations based on topic and student level to guide study plans.

import_user_document

Upload user-provided documents to be analyzed and integrated into learning resources and plans.

generate_learning_plan

Create a structured learning plan with clear objectives and suggested resources tailored to the learner.

extract_article_key_points

Extract key points, methodologies, and findings from medical articles to support quick review and exam prep.