InterSystems Developer Community

10/13/2026

🤖 New programming contest: build an Interoperability workflow where an AI Hub agent handles a decision-making task.

💰 Compete for a share of $12K prize pool
📅 Submission period: October 26 - November 8, 2026

👉 https://community.intersystems.com/post/intersystems-programming-contest-ai-agents-interoperability-workflows

10/07/2026

💡 Five developer challenges are now open in the Community Bounty Program.

🎁 Build an app and earn 10K+ points redeemable for gifts. Challenges cover AI datasets, FHIR, WhatsApp & Supabase.

📅 October 31, 2026

👉 https://community.intersystems.com/post/community-bounty-program-idea-application-%E2%80%94-round-3-live

10/06/2026

🔄 HL7 routing in can require more context than the incoming message provides.

A practical way to handle this is to enrich the message before the Business Rule evaluates it. This walkthrough shows how a BPL Business Process can extract values from an incoming HL7 message, store them in a %DynamicObject, and make that metadata available to downstream routing logic without modifying the original message.

The guide also covers some important implementation details, including why BPL context properties must be declared in advance, how to access dynamic object properties safely, and how to verify the flow in Message Viewer and Visual Trace.

🔎 Read the full article with ObjectScript snippets, screenshots, and the complete IRIS Interoperability production setup by Muhammad Waseem:

https://community.intersystems.com/post/business-rules-deep-dive-dynamic-objects-and-property-access-pitfalls-part-1

09/30/2026

🤖 Moving AI agents from pilot to production requires more than access to a language model.

Once agents need enterprise data, business logic, permissions, external tools, and human oversight, the surrounding infrastructure becomes just as important as the model itself.

AI Hub brings these capabilities into one platform, providing an integrated environment for building, deploying, and managing AI agents and agentic workflows.
Native Model Context Protocol support is bidirectional: IRIS data and business logic can be exposed as governed tools, while AI Hub agents can connect to third-party MCP servers and enterprise applications. AI Hub also uses IRIS capabilities for security, interoperability, and human-in-the-loop workflows, allowing agents to run alongside existing applications and development processes.

🔎 Read more about AI Hub, governance, and production agentic workflows on the Developer Community: https://community.intersystems.com/post/intersystems-iris-ai-hub-build-connect-and-govern-ai-agents

09/28/2026

🗳️ The contest projects are in — now it’s time to choose your favorites.

Explore the submissions for “Build Your Own Management Portal” programming contest and see how developers approached management tasks through their apps and interfaces.

Support the solutions you like most in the Community Nomination and follow the daily leaderboard updates on the Developer Community.

📅 Voting: September 28 - October 4, 2026
👉 Explore the apps and cast your vote: https://community.intersystems.com/post/time-vote-intersystems-programming-contest-build-your-own-management-portal

09/25/2026

⏳ The deadline is approaching for the InterSystems Programming Contest: Build Your Own Management Portal.

Now is the time to finish your entry – or jump in while submissions are still open! 😉

The challenge: Develop a GUI powered by management APIs and bring your own approach to the Management Portal experience.

💰 $12,000 prize pool
🏆 Experts, Community, and Freshmen awards are waiting for their winners.
📅 Deadline: September 27, 2026

👉 Get your project ready and submit: https://community.intersystems.com/post/intersystems-programming-contest-build-your-own-management-portal

09/24/2026

⚙️ A simple CI workflow can automate InterSystems Health Connect deployments with GitHub and Jenkins

This implementation uses GitHub as the source repository and Jenkins as the deployment orchestrator. Jenkins connects to the development server over SSH, while Linux and ObjectScript scripts identify changed files, load and compile them into IRIS for Health, and restart the interoperability production.

The workflow supports incremental deployment by comparing Git commits and processing only added, modified, or renamed files. The setup provides a straightforward foundation for bringing InterSystems Health Connect development into a Git-based continuous integration process.

🔎 Read the full implementation, scripts, and deployment flow on the Developer Community: https://community.intersystems.com/post/continuous-integration-iris-git-and-jenkins

IntroductionIn healthcare interoperability environments, InterSystems Health Connect typically contains critical components such as productions, business processes, operations, services, utility classes, routines, and other ObjectScript artifacts.

09/22/2026

🔎 FHIR APIs are powerful, but many real clinical questions require more than a single search.

EvidenceQL is an application that introduces an evidence-first query runtime between the AI agent and the FHIR backend. Instead of letting an LLM make arbitrary FHIR REST calls, the agent produces a structured FHIRQL JSON plan that is validated, compiled, executed, and returned together with the original FHIR resources used as evidence.

This makes it possible to handle healthcare data questions that go beyond a single FHIR search, including:

🔹 multi-resource cohort queries
🔹 temporal windows and trend detection
🔹 missing follow-up and care-gap analysis
🔹 medication overlap
🔹 patient evidence timelines
🔹 traceable explanations of why a result matched

The runtime also exposes an MCP interface for AI agents, with a controlled workflow:

👉 capabilities → validate → compile → execute → explain

The key idea is evidence-first ex*****on: results remain tied to the underlying FHIR resources, values, warnings, and ex*****on metadata rather than relying on the model to generate an unsupported clinical conclusion.

📖 Read the full architecture with code examples on the Developer Community: https://community.intersystems.com/post/evidenceql-evidence-first-query-runtime-ai-agents-over-fhir-apis

09/21/2026

🤖 How well do LLMs perform on ObjectScript and ?

The new ObjectScript LLM Benchmark is the first open benchmark focused specifically on LLM competence with ObjectScript and the IRIS platform. It evaluates 14 models on 140 hand-written tasks across code generation, code comprehension, and bug fixing. For 93 tasks, generated code is compiled and executed directly on a live IRIS instance.

Some key findings:
🔹 Overall accuracy ranges from 0.515 to 0.851
🔹 The top nine models have overlapping 95% confidence intervals, so the current benchmark cannot reliably rank them against one another
🔹 Code generation is the hardest and most discriminating task family
🔹 Model cost and ObjectScript performance do not move together consistently: similarly accurate models can differ substantially in inference cost

The benchmark offers a practical look at LLM performance for ObjectScript and IRIS development.

Read the full article on the Developer Community: https://community.intersystems.com/post/objectscript-llm-benchmark-ex*****on-graded-results-14-models

We've published the first open benchmark measuring LLM competence on InterSystems ObjectScript and the IRIS platform: [ObjectScript LLM Benchmark](https://github.com/isc-nmitchko/ObjectScript-LLM-Benchmark). # # Summary The [ObjectScript LLM

09/18/2026

🧑‍💻 Show your competitive spirit. Design an experience around your development workflow.

The new Programming Contest: Build Your Own Management Portal is gaining speed. Your challenge is to create a GUI powered by management APIs for core management tasks: from APIs and permissions to security, system resources, and logs.

The technical scope is set, but you shape the experience. Now is the time to get creative and join the contest.

💰 $12,000 prize pool
📅 Submit by September 27, 2026
👉 Start here:

Hi Community,We’re excited to announce our next online programming contest:🏆 InterSystems Programming Contest: Build Your Own Management Portal 🏆Duration: September 14 - October 4, 2026Prize pool: $12,000🌟 New this round: we’re introducing the