Taiwan-Health

Taiwan-Health MCP Server provides a comprehensive Taiwan-localized data integration platform that exposes a range of healthcare information through MCP tools. It aggregates ICD-10 diagnostic and procedure codes, Taiwan FDA drug data, health foods, nutrition, LOINC lab codes, and clinical guidelines, and exposes these via FHIR R4 resources such as Condition, Medication, and MedicationKnowledge. The server offers tooling to search ICD-10-CM and ICD-10-PCS, convert findings into FHIR resources, query integrated drug and health food data, retrieve nutrition and LOINC information, and access clinical guidelines and care pathways. To use the server, clone the repository, install dependencies, prepare data, and run the Python server; then interact with the endpoints using standard FHIR/REST calls or the provided MCP tooling described in the docs.

Prerequisites:

• Python 3.8 or higher
• Git
• Network access to download dependencies and data files

1. Clone the repository git clone https://github.com/healthymind-tech/Taiwan-Health-MCP.git cd Taiwan-Health-MCP

2. Create and activate a Python environment (optional but recommended) python -m venv venv source venv/bin/activate # on Windows use venv\Scripts\activate

3. Install dependencies pip install -r requirements.txt

4. Prepare data

   • Download the ICD-10 Excel files and place them into the data/ directory as instructed in the docs

5. Run the MCP server python src/server.py

Optional: Docker (recommended for deployment)

1. Ensure docker and docker-compose are installed
2. Use the project’s docker-compose setup as documented in the README docker-compose up -d

Prerequisites summary

• Python 3.8+ (or a containerized environment)
• Dependency management (pip) and access to data sources
• Data preparation steps as described in the docs

Tips and notes:

• The MCP server emphasizes Taiwan-localized data and supports FHIR R4 representations. Be aware that certain advanced FHIR validation and terminology bindings may not be fully implemented; use an external FHIR validator for production deployments.
• Health foods and dietary suggestions are in development and should be reviewed by qualified professionals before clinical use.
• If running in production, consider configuring environment variables and proper data synchronization scheduling as described in the docs (e.g., automatic data refresh for ICD-10, FDA datasets, and LOINC mappings).
• For debugging, enable unbuffered Python output (set PYTHONUNBUFFERED=1) to get real-time logs.
• Refer to the full docs site and local MkDocs docs for API details, resource mappings, and example queries.