Amazing power of HL7 FHIR in agentic era

As mentioned in my previous blog dated in Jan this year,  I started to explore how to implement production quality FHIR API and Data Platform without single line of coding. During the past 3 months during my spare time (btw I am also out of office for few months as I did surgery last Nov and on recovery for few months afterwards), but whenever there is few minutes, I prompted AI to plan out the spec, implementation plan and generate code.

Now the the API and Data Platform is fully ready, all is written by AI (Claude Code) with these major features

1. It provides out of box HL7 FHIR R4B/R5 resources
2. It supports custom resources
3. It provides set of skills to define custom resources, define FHIR profiles etc
4. It supports publish/subscribe notification
5. It is equipped with BDD to do full end-to-end functional tests
6. It is also able to deploy to local Docker environment and AWS Cloud Environment where all the AWS services are configured as private  and integrated through VPC endpoint.
7. It provides out of box integration through MCP server with any AI agent as described below. It only requires 3 MCP tools (Pls see the demo at https://youtu.be/pM-RERPoqrQ) 

Go check out the source code on GitHub, I welcome passionate contributors to make this open source product more powerful and more robust.

https://github.com/sg-victorchai/fhir4java-agents