Developer builds offline Android health dashboard using local LLMs

By Splitfeed EditorialLast updated Sep 26, 2026, 19:06 UTC

On September 25, 2026, a developer posted a personal Android project called Hamilton to Hacker News under the title "Show HN: Hamilton – a no-internet Android health dashboard, named after my dog." The app tackles a specific friction point in consumer fitness technology: when a user owns devices and apps from multiple manufacturers, those products rarely communicate with each other, leaving health metrics scattered across separate interfaces or counted more than once. Hamilton pulls that data into a single dashboard and uses local large language models to process it, meaning no data leaves the device and no internet connection is required at any point.

The developer's starting point was a different idea entirely, an app that would have disabled Google Pay whenever the phone came within 50 metres of a McDonald's, before they pivoted to the health dashboard. That detail is minor on its own, but it illustrates a broader pattern in the current LLM moment: individuals are now prototyping several distinct app concepts in the time it once took to build one, and the projects that survive are often solutions to problems the developer personally lives with rather than products aimed at a market. Hamilton was shared publicly but without any mention of download figures, funding, or outside collaborators, placing it firmly in the personal-tool category. The genuinely contested question is whether local LLMs are yet capable enough to add real analytical value to health data, or whether they are being used here primarily as a convenient processing layer whose outputs remain to be scrutinised.

With no published reactions from the Pro-AI, Anti-AI, or Middle Ground camps at this stage, their positions can only be anticipated based on how each typically responds to stories of this kind. The Pro-AI camp would likely point to Hamilton as evidence that local AI is now accessible enough for solo developers to build privacy-preserving alternatives to cloud-dependent health platforms. The Anti-AI camp would probably raise questions about whether LLM outputs are reliable enough for health contexts and whether the framing of "AI-powered" adds meaningful capability or is largely cosmetic. The Middle Ground camp would typically argue that the privacy architecture is genuinely valuable, while pressing for transparency about what the local model actually does with the aggregated data and how its conclusions should be weighted.

The argument would sharpen considerably if the developer publishes technical detail about which local model is used, what queries or inferences it runs on the health data, and how the aggregation logic handles conflicts between manufacturer data formats. Any of those disclosures would give critics and supporters concrete material to assess rather than a concept.