Personal Health Dashboard

I'm taking five fewer supplements than I was last month, and the reason isn't a podcast or a YouTube video. It's that I finally let Claude read every blood test I've taken for the last seven years.

For most of those seven years, my supplement stack worked the way most people's do. A friend mentioned something. A podcast guest swore by it. A protocol on Instagram looked compelling. I'd add it, take it for six months, never test whether it was actually doing anything for my body specifically. By the time I stopped to count, I was on a daily stack that cost about $700 a month and was, as far as I could tell, mostly running on faith.

Then I built this.

What it is: a private Claude project loaded with seven years of my own bloodwork, my full supplement and peptide stack, my baseline goals, and a system prompt that turns Claude into something between a research analyst and a functional medicine practitioner. Not medical advice. Not a replacement for a doctor. The analysis layer between "raw numbers I don't fully understand" and "decisions I'm confident enough to act on."

The first time I ran the supplement audit, Claude flagged 5 supplements for outright drop and 9 more for review. None of them came as a shock. All of them came with reasoning I could verify against my actual data. Things I'd been taking out of habit, not biology.

This is the two-piece build. The foundation (chat-based audit, takes 30 minutes to set up). The upgrade (visual dashboard you re-open anytime, takes an extra hour). Both prompts are below. Run as much of it as you want.

What you'll need

• A Claude Pro or Max subscription. Free tier won't handle the file uploads or project memory you need for this.
• Your bloodwork results. PDFs from Quest, LabCorp, Function Health, InsideTracker, your annual physical, anywhere. The more years the better. Even one or two years works. I had seven and the patterns over time were where the real value showed up.
• Your current supplement and peptide stack. Names, doses, when you started each one, why you started.
• Optional but worth doing: wearable data exports (Whoop, Oura, Apple Health), sleep tracking, anything quantified you can dump in.

Privacy note: a Claude project is private to you. Nothing in it gets used to train models. Don't paste your bloodwork into a shared chat or a public artifact, but a project is the right home.

Piece 1: The foundation (the 95% piece)

This is the part most people don't need to upgrade past. A chat, a system prompt, and a few starter prompts that surface decisions you can actually act on.

Setup

1. Open Claude. Create a new project. Call it "Personal Health Analyst" or whatever you want.
2. Upload all your bloodwork PDFs into the project knowledge.
3. Add a text file or note with your current supplement stack, doses, start dates, and stated reasons for each.
4. Add a second text file with your current health goals (e.g., "longevity, performance, body composition, sleep quality").
5. Paste this as the project's system prompt:

You are my personal health analyst. You have access to my full bloodwork history, my current supplement and peptide stack, and my stated health goals.

Your job is to be the analytical layer between my data and my decisions. You are not a doctor. You don't replace one. But you are the equivalent of a research analyst who has spent the weekend going through my actual numbers and is now ready to brief me.

When I ask you questions:
- Cite the specific bloodwork values you're reasoning from, with dates
- Show your reasoning, not just conclusions
- Flag what you are uncertain about
- Tell me when something requires a real doctor, not you
- Surface patterns I might not see (e.g., values trending in the wrong direction over multiple years, even within "normal" range)
- For supplement questions, tell me whether my data supports the supplement, contradicts it, or is silent on it. "Silent" is a valid answer.
- Never recommend dropping a prescribed medication without my doctor.

Be direct. Be specific. Don't hedge to protect yourself. I want the analysis a smart friend who happens to be a research analyst would give me, not a watered-down version designed to avoid liability.

That's the foundation. Five minutes of setup, and you now have a private analyst that knows your actual body.

Prompt 1: The supplement audit (the killer prompt)

This is the one that flagged 5 of my supplements for drop.

Go through my entire current supplement and peptide stack. For each one:

1. What's the biological case for me specifically taking this, based on my bloodwork?
2. Is there a measurable marker in my data that this supplement is moving in the right direction?
3. Is there a marker it should be moving that isn't?
4. Are there any markers it might be hurting?
5. Final call: keep, reduce, drop, or needs more data?

Be brutal. I'd rather drop something that's working slightly than keep five things that aren't doing anything measurable.

The output of this prompt is what changed my month. Take it seriously.

Prompt 2: The trend scanner

This is what most doctors don't do because they don't have the time.

Look at my bloodwork over all the years I've uploaded. Find every marker that is:

1. Trending in the wrong direction year over year, even if all values are still in the "normal" range
2. Trending in the right direction (so I know what's working)
3. Outside the optimal range used by functional medicine practitioners (not just the standard lab range)
4. Showing volatility I should investigate

For each one, tell me what it might mean and what would be worth testing next.

This is where the years of data start paying off. Single bloodwork results are snapshots. Multi-year trends are stories. Most of the actionable signal is in the stories.

Prompt 3: The next test

Use this when you're about to do your next round of testing.

Based on my full data and goals, what should I be testing in my next round of bloodwork that I haven't tested before? What patterns in my data suggest a gap I should fill?

This has gotten me to add tests my regular panel never covered. Cheaper to test than to guess.

Piece 2: The upgrade (the visual dashboard)

This part is optional. Most people get 95% of the value from the chat above. But if you want the dashboard view that you can re-open anytime, paste this prompt into a fresh conversation inside the same project.

You don't need to understand the technical part. Claude builds it. You just paste, accept, and use.

Based on everything in this project (my bloodwork, supplements, goals), build me a single-page HTML dashboard artifact that visualizes my health data over time. The dashboard should include:

1. A top-row summary panel: my key markers as of the latest test, color-coded against optimal ranges (not just standard ranges)
2. A trends section: line charts of my most important markers over time (lipid panel, metabolic markers, inflammation markers, hormones, vitamins/minerals, anything you've identified as trending in the wrong direction)
3. A supplement panel: my current stack, with a "supported by data / contradicted by data / silent" tag on each one based on the audit
4. A "watch list" panel: markers you've flagged as needing attention
5. An "ask Claude" panel at the bottom with three pre-written follow-up questions I should be asking about my data

Style: dark mode, clinical, minimal. No motivational quotes. No emojis. Clean visual design. Designed for desktop but readable on mobile.

Save the artifact. Re-open anytime. When you get new bloodwork, drop it into the project and ask Claude to regenerate the dashboard with the new data factored in.

What this is not

A few things worth being clear about, because the medical and supplement space is full of people who aren't:

• This is not medical advice. Claude is good. It is not a doctor. If something serious surfaces, take it to a doctor.
• Don't drop a prescribed medication based on this. Ever.
• Don't drop a supplement just because mine got dropped. My biology is not yours. The point is to run this against your data, not to copy my conclusions.
• The "optimal" ranges Claude uses aren't universally agreed upon. Functional medicine practitioners use tighter ranges than standard labs. Both are defensible. Ask Claude to explain which range it's using and why.
• Bloodwork is a snapshot. Wearable data, symptoms, energy, performance — all of it matters. This pipeline is one input, not the whole picture.

Why this changes how you spend money

The supplement industry runs on hope. You take something because someone you trust said it works for them. You take it for six months. You never test whether it's actually doing anything in your body specifically. Repeat across ten supplements and you're spending $700 a month on a stack you can't defend with a single number from your own labs.

This pipeline kills that pattern. Not by telling you to stop taking supplements. By giving you the analytical layer that finally tells you which ones are working for your body and which ones are running on faith.

What changed for me wasn't the supplements I dropped. It was that I now make every health decision against the same baseline: what does my actual data say. Not what Bryan Johnson says. Not what the longevity podcast guest says. Not what the supplement company says. What seven years of my own blood says.

That shift is the real upgrade. The dashboard is just the surface.

What this changes

This is the build I run on myself. Two prompts you can copy tonight. One that works for anyone with a single year of bloodwork. One that gets sharper the more data you load.

The pattern beyond health: any domain where you've been making decisions on stories instead of your own data, this same structure applies. Investing. Content strategy. Sleep. Training. Pick the domain. Load the data. Build the analyst. Make decisions against your actual numbers instead of someone else's protocol.

The people who run themselves like this in the next decade are going to look very different from the people who don't.