Introducing Eka MCP Server: India’s AI-First Medical Knowledge Bridge for LLMs

Enhancing AI accuracy in healthcare with curated medical knowledge

At Eka Care, we envision a future where every doctor has evidence-based AI support tailored to Indian medical practice—no guesswork, just clarity. 

Eka Care is excited to announce the launch of Eka MCP Server, our open-source MCP server designed to enhance the accuracy and reliability of AI assistance in healthcare settings.

With the Eka MCP Server, we’re not just reducing hallucinations—we’re redefining how AI safely interacts with clinical decision-making. This positions Eka Care as the AI assistant doctors can trust in Indian healthcare.

Built on the Model Context Protocol (MCP) - an open standard that allows Large Language Model (LLM) clients (like Claude Desktop, ChatGPT) to extend their capabilities by providing additional tools and resources. The Eka MCP Server addresses healthcare challenges by providing verified medication information, and standardized treatment protocols from authoritative medical publishers. 

This collection of tools addresses the issue of accessing critical medical information without disrupting patient care workflows.

Challenge of information access vs. patient focus

Healthcare professionals frequently need to switch context to access additional information while treating patients. Toggling between patient care and information retrieval disrupts workflow and reduces efficiency. The information is also present in disparate sources, often away from the EMR. LLMs have proven to serve as a bridge to provide this information, however, there's an inherent risk of hallucination or inaccuracies. In healthcare settings, this cannot be tolerated.

Here's an illustration of LLM’s lack of localised information leading to wrong answers, when asked about zin 10mg, ChatGPT thinks it’s a zinc supplement. Looking at 1mg and PharmEasy, popular medicine retailers in India, Zin 10 mg is actually an anti-histamine for allergies containing Cetirizine.

‍

Using the Drug Search tool available on Eka MCP Server, LLMs can get details about a drug, search through Eka Care’s drug database of medicines with the drug name or generic composition. Shown below is an example of the same query about zin 10 mg done with Claude with MCP.

The Eka MCP Server represents a significant advancement in how healthcare professionals can interact with LLM. By grounding LLM responses in curated medical information we're creating a more reliable and seamless experience that healthcare providers can trust.

About the technology

The Eka MCP Server is built on the Model Context Protocol (MCP), an emerging open standard that enables AI models to access external tools and information sources. This technology allows AI assistants like Claude to ground their responses in verified information rather than relying solely on their training data.

How Eka MCP Server transforms healthcare AI

The Eka MCP Server brings several groundbreaking capabilities to healthcare professionals:

• Medical Accuracy: Grounds AI responses in verified healthcare information from Eka Care's extensive databases
• Seamless Workflow: Enables healthcare professionals to access critical information without disrupting patient care
• Reduced Hallucinations: Relies on curated medical data rather than AI's general knowledge
• Open Ecosystem: Functions as part of the growing Model Context Protocol (MCP) open standard, ensuring compatibility with various AI platforms

What makes this solution particularly powerful is its ability to serve as a bridge between large language models (LLMs) like Claude and verified medical information sources. This ensures that AI assistants provide healthcare professionals with accurate, reliable information at the moment of need.

Specialized MCP tools for Healthcare

Eka MCP server launches with two tool suites specifically designed for healthcare professionals

Indian branded drug search

Eka Care has a curated dataset of 500,000+ branded drugs in India along with its form, generic composition and manufacturer. Contextually, LLMs can invoke this tool to understand more about a branded drug or query the database with drug name or even the generic composition. Augmenting generic information thus extends the capability of existing SOTA LLMs.

Indian treatment protocol search

Quite often, doctors refer to standard guidelines of treatment published by eminent medical bodies. These are used for general guidelines, nuanced information in case of comorbidities. Through this tool, doctors can access information that is otherwise locked up in PDF documents. 

While LLMs can answer questions about international clinical guidelines, they struggle and hallucinate with Indian guidelines which have nuances of treating Indian patients.

• Access to standardized guidelines, procedures, and decision pathways published by authoritative medical bodies
• Current coverage includes 175 medical conditions/tags, 180 treatment protocols, and multiple authoritative publishers

‍

Here’s a sample conversation on Claude showcasing this tool’s capabilities - https://claude.ai/share/eb228805-41a0-4bef-8f50-4ae6bb90bb3a

Click on the tool invocation toggle to see the tool input and tool response

Integration and Compatibility

The Eka MCP Server integrates seamlessly with

• Any other platform that supports the MCP specification
• Anthropic's Claude Desktop Client, which have native MCP support
• Cursor and Windsurf applications

The server code is open-source and available now on GitHub and PyPI, making it accessible to healthcare organizations of all sizes. While the MCP server code itself is open-source, access to Eka Care’s comprehensive medical APIs requires valid credentials, ensuring proper validation and responsible use of the medical information.

Get started with Eka MCP Server

Detailed instructions available on the github repository - https://github.com/eka-care/eka_mcp_server

Here's how to get started:

1. Obtain your developer key by reaching out to ekaconnect@eka.care
2. Follow the simple installation instructions available on our GitHub repository here
3. Start using the MCP server on supported clients like Claude desktop

Roadmap

• Our drug database is expanding and we try our best to have a rich catalogue
• The treatment protocols will further be increased to cover wider range of conditions
• Soon, we are planning on launching another tool to provide ICD10, SNOMED-CT linking for any given text by segmenting entities and linking them to specified coding standards.

Join the healthcare AI revolution

The fusion of artificial intelligence with verified medical knowledge represents the future of healthcare assistance. By adopting the Eka MCP Server, healthcare organizations can ensure their AI tools provide accurate information while maintaining the critical focus on patient care.

For more information or to start using Eka MCP Server today, visit GitHub repository, developer.eka.care or contact us ekaconnect@eka.care‍

‍