You’ve been there. You open ChatGPT, paste your latest blood work, describe that weird symptom you’ve had for three months, and ask for insights. The AI gives you a thoughtful, nuanced answer. You feel a moment of hope. Then you close the tab, and the next time you have a question, you have to re-type everything. The context is gone. The magic evaporates.
This is the dirty secret of AI for personal health: the models are smart enough, but your data is a mess. Every query is a fresh start, and every fresh start costs you five minutes of digging through test results, remembering dates, and typing out the same story. It’s exhausting. And it’s why most people give up after a few tries.
The bottleneck isn’t AI’s reasoning. It’s your data’s messiness.
I saw this firsthand when I tried to get ChatGPT to help me with my chronic migraines. I’d list triggers, medications, sleep patterns, and then ask for a pattern analysis. The answer was great. But the next week, I needed to ask a follow-up about a new medication interaction. I had to start from scratch. I couldn’t just say “remember my timeline?” because the AI doesn’t remember anything. It’s a fresh conversation every time.
That’s where a new kind of tool comes in. Not an AI wrapper. Not another chatbot. A personal health data hygiene tool that builds a reusable, structured timeline of your health events. You log your migraines, your test results, your medications — all in natural language, with fuzzy dates. It lives on your phone, locally, no account required. Then, when you want to ask an AI for advice, you export that timeline as a prompt. The AI gets your full context. Every time. No re-typing. No forgotten details.
This is what traditional healthcare EHRs have failed to do for patients: give you ownership and portability of your own medical narrative.
The app is called Kurozu, and it’s deceptively simple. It’s a timeline of events and facts. You add events in natural language (“Had a migraine on March 15, triggered by a storm, took sumatriptan, worked after 2 hours”). It syncs with Apple Health for basic insights. And it’s local-first — no cloud, no privacy risk. The real product isn’t the AI at all. It’s the discipline of building your own health timeline. Once you have it, you can paste it into any AI, any tool, any doctor. You become the owner of your story.
Every AI health query without a context timeline is like asking a genius to solve a puzzle without showing them the pieces.
The skeptic will say: “But I can already do this with a notes app.” True. But you won’t. The friction of manually structuring data for AI consumption is the barrier. Kurozu reduces that friction to a single tap. It’s the difference between a messy pile of receipts and a tax-ready spreadsheet. The app forces you to be consistent, to date things, to categorize. That structure is what makes the AI useful.
And the contrarian take? The real value here isn’t the AI advice at all. It’s the act of building the timeline. In doing so, you start to see your own patterns. You become more aware of your health. The AI is just a bonus. The tool is actually a mirror.
So before you ask another AI for health advice, do this one thing: build your context. Don’t make the AI guess. Give it the full picture. The future of personalized health isn’t smarter models — it’s reusable context. And that starts with you.
FAQ
Q: What makes this different from just using a notes app or a simple text file?
A: A notes app doesn't enforce structure — you'd still have to manually format your data for each AI query. Kurozu forces consistency, fuzzy date handling, and one-tap export to a prompt that any AI understands. It's the difference between a messy pile of receipts and a tax-ready spreadsheet.
Q: Does this actually help the AI give better advice, or is it just a convenience thing?
A: It dramatically improves quality. LLMs are context-limited. A full timeline gives the AI your medication history, symptom patterns, triggers, and lab results in one shot. Without it, the AI is guessing based on a single, narrow question. The difference is analogous to a doctor who has your full chart versus one who only sees your current complaint.
Q: Isn't it safer to just let AI access my Apple Health directly? Why build a separate timeline?
A: Apple Health data is sparse and not structured for narrative health queries. It tracks steps and heart rate, not 'I had a migraine after eating cheese.' A timeline lets you add subjective, qualitative data that AI can actually reason about. Plus, local-first means no third-party cloud — your data stays on your phone, not on some AI company's server.