HN
Today

Show HN: GlycemicGPT – Open-source AI-powered diabetes management

GlycemicGPT is an open-source, self-hosted AI platform designed by a Type 1 diabetic engineer to empower users with advanced glucose monitoring and insulin pump data analysis. It offers personalized AI insights, conversational chat with clinical knowledge, and predictive alerts, all while ensuring user data privacy by running on personal infrastructure. This project appeals to the HN crowd for its blend of cutting-edge AI, vital health tech, and a strong commitment to open-source principles and user control.

18
Score
3
Comments
#2
Highest Rank
3h
on Front Page
First Seen
May 15, 9:00 AM
Last Seen
May 15, 11:00 AM
Rank Over Time
257

The Lowdown

GlycemicGPT presents itself as a robust, open-source AI-powered platform for diabetes management, born from a Type 1 diabetic software engineer's personal need for better data insights. This project aims to bridge the gap in clinical review and data analysis by providing users with a comprehensive, self-hosted system to monitor and understand their glycemic patterns.

  • Data Integration: The platform connects directly to popular devices such as Dexcom G7 (via cloud API) and Tandem t:slim X2/Mobi insulin pumps (via direct BLE), with future plans for Nightscout integration.
  • AI-Powered Analysis: It offers daily briefs summarizing glucose patterns, meal response analysis, and a conversational AI chat backed by clinical knowledge. Users can 'Bring Your Own AI' (BYOAI) providers like Ollama for local processing or cloud services such as Claude and OpenAI.
  • Key Principles & Safety: GlycemicGPT is strictly for monitoring and analysis; it does not control medical devices or deliver insulin. It is self-hosted via Docker or Kubernetes, ensuring data privacy and preventing vendor lock-in. Crucially, it includes extensive disclaimers, emphasizing that it's experimental alpha software and not a replacement for professional medical advice.
  • Technical Stack: The system is built with a modern stack including FastAPI (Python) for the backend, Next.js (React) for the web dashboard, and Kotlin for Android and Wear OS apps, featuring a plugin SDK for device extensibility.
  • Community & Licensing: Released under the GPL-3.0 license, the project actively seeks contributors, especially those with BLE/Android experience or expertise in diabetes technology.

Ultimately, GlycemicGPT empowers individuals to take a proactive, data-driven approach to their diabetes management, offering advanced tools and insights while firmly placing the user in control of their data and AI processing, albeit with critical warnings about its supplementary role to professional medical care.