I wish there was more options for export, I was especially looking for breathing disturbances (sleep apnea prediction) but I’m sure it a bunch of other stats would be useful.
I'm a runner with years of data sitting in Apple Health. Paces, heart rate, distance, elevation, all of it. I wanted to feed this into an LLM to help me build a personal training plan, analyse some trends, and spot patterns across the season.
Seemed like this should be straightforward. But it's not. Apple's built-in export gives you a giant XML blob that chokes most tools, iff you're able to export the data in the first place (I stopped my export after 6 and a half hours). Third-party apps either want a subscription or lock you into their own dashboards. All I wanted was clean, structured data I could paste into Claude.
So my wife and I built Vitalina (https://vitalina.app). It's an iOS app that exports Apple Health metric as CSV, JSON and PDF. Workouts, heart rate, sleep, whatever. Pick a date range, pick your metrics, export. The output is compact and structured to actually be useful for LLM analysis.
No account, no cloud, everything stays on-device until you decide where the file goes.
If you've been trying to use LLMs with your own health data, I'd love to hear what you're doing with it and what would make this more useful.
I wish there was more options for export, I was especially looking for breathing disturbances (sleep apnea prediction) but I’m sure it a bunch of other stats would be useful.
I'm a runner with years of data sitting in Apple Health. Paces, heart rate, distance, elevation, all of it. I wanted to feed this into an LLM to help me build a personal training plan, analyse some trends, and spot patterns across the season.
Seemed like this should be straightforward. But it's not. Apple's built-in export gives you a giant XML blob that chokes most tools, iff you're able to export the data in the first place (I stopped my export after 6 and a half hours). Third-party apps either want a subscription or lock you into their own dashboards. All I wanted was clean, structured data I could paste into Claude.
So my wife and I built Vitalina (https://vitalina.app). It's an iOS app that exports Apple Health metric as CSV, JSON and PDF. Workouts, heart rate, sleep, whatever. Pick a date range, pick your metrics, export. The output is compact and structured to actually be useful for LLM analysis.
No account, no cloud, everything stays on-device until you decide where the file goes.
If you've been trying to use LLMs with your own health data, I'd love to hear what you're doing with it and what would make this more useful.