Context-Aware LLMs, a very powerful way to integrate LLMs into DHIS2

Hello everyone, my name is Abdelatif Abualya, I am a developer at HISP MENA. I want to tell me you about a new approach to integrate generative AI into DHIS2

Introducing the DHIS2 AI Assistant – A Context-Aware Chrome Extension for DHIS2

I’ve been working on a Chrome extension that brings AI-powered, context-aware assistance directly into any DHIS2 instance.

What makes it different? Unlike generic chatbots, this assistant automatically knows where you are in DHIS2 — which program, facility, or stage you’re viewing — without you having to explain anything. It monitors URL changes in real time and resolves full program metadata behind the scenes, so conversations feel natural and instant.

Key capabilities:

16 specialized DHIS2 tools covering analytics, tracked entity queries, event data, visualizations, maps, and more.

Autonomous metadata creation — describe a new program in plain language, and the AI builds the full dependency chain using an atomic, validated workflow.

Natural language analytics— ask “How many patients completed this stage?” or “Show ANC trends as a chart,” and the AI handles all API calls automatically.

Agentic reasoning loop with up to 12 tool call iterations per query.

I just want to note that I haven’t officially published the extension yet, as I am still testing the extension, and working on ways to improve it.

Here is a short demo:

6 Likes