# I really wish I could connect WHOOP, Function Health, Diet, and Supplements

By Morgan Linton · Published 2026-01-08 · X No. 2 · 1,125 words · https://blog.morganlinton.com/on-x/i-really-wish-i-could-connect-whoop-function-health-diet-and-supplements/

![](../../assets/on-x/G-J-0DEa0AAKaKG.jpg)

Okay, this is going to be a pretty informal article, so I guess you can also feel good knowing it's not written by AI. I wrote a comment on OpenAI's announcement of ChatGPT Health yesterday that ended up getting a lot more attention than I had expected, so I thought - why not dive a little deeper?

First, here's what I wrote:

> Wild. I have literally been working on building an app to track all of my health data.
>
> Curious if this can pull in data from WHOOP and Function Health and cross reference?
>
> That's the main thing that's been missing for me is connecting blood test results with sleep, exercise,
>
> — [@morganlinton on X](https://x.com/morganlinton/status/2008993736622330367)

As I've talked with more and more people who know about what OpenAI is doing with ChatGPT Health, it does sound like my dream isn't going to come true here. And I get it - a lot of people are using AI ChatBots to talk about health, and it makes a ton of sense for OpenAI to launch an AI ChatBot completely focused on health.

I'm in a smaller group of people that has gone a bit nuts with all the different health tech out there. I use [@WHOOP](https://x.com/@WHOOP) to monitor my heart, exercise and recovery, but it's the heart monitoring that has kept me a loyal customer for 5+ years. I use [@function](https://x.com/@function) for regular blood testing and advice on diet and supplements.

The challenge I have is that WHOOP gives me insights on things I should change about my lifestyle, so does Function, but the two aren't connected, and neither one of them tracks if I actually take this advice, or if I do, if I'm following it correctly.

I'll give you an example of where I could use some help, and I guess other people could too based on the response to my tweet from yesterday. Function tells me I need to eat more Gooseberry, Dark Chocolate, Horseradish, Wasbi, and Eggs.

![](../../assets/on-x/G-KAnnPbQAAzMTk.png)

They also have a list of supplements I should start taking, with my top five being Omega 3, Plant Sterols, Fiber Supplements, Coenzyme Q10, and Berberine. Okay, got it 🫡

![](../../assets/on-x/G-KA1wOa0AMjzPa.png)

WHOOP then gives me insights like this each day:

![](../../assets/on-x/G-KBn6Aa0AEm-hX.jpg)

But WHOOP doesn't know anything about my blood test results from Function, about the diet changes I've made, or the supplements I'm taking. Function knows a little bit about how much I sleep, and how many steps I take since it integrates with Apple Health, but WHOOP only shares a fraction of its data through Apple Health.

Here's some of the data that WHOOP doesn't report to Apple Health:

- **HRV (Heart Rate Variability)**
- **Continuous, all-day heart-rate timeline**

- **WHOOP’s proprietary scores** like **Recovery**, **Strain**, **Stress**, etc.
- **Sleep stages**

So I've found myself in an interesting situation. I've changed my diet, added supplements, and feel like I'm in a good groove. What I'm missing is a way to actually track if what I'm doing is working, for me, and if I really need to take all those supplements, or augment my diet as much as I have.

The feedback loop is a bit broken. Sure I get my Function Labs done every six months, but I find they're just telling me the same thing every six months with small tweaks. My Vitamin D and Omega-3's are improving, but some of my heart metrics which are impacted by things like genetics, exercise, sleep, etc. aren't being analyzed quite right IMHO.

Of course, I'm not a cardiologist, so what do I know. What I am doing is going to a cardiologist later this month to see what they think of all of this data. But I run the risk of a cardiologist who has been doing things the same way forever and things all this "new fangled" health tech stuff is just creating more paranoia than good. Which could mean he'll just quickly scan the data and say, "you're fine, just keep eating a Mediterranean diet, keep exercising like you do, etc."

My guess is - if there was a way to actually track what I'm eating, the exact supplements I'm taking every day (and the doses), and connect this with my WHOOP data, and Function results, I would get a lot more personalized, actionable insights.

I've spent countless hours chatting with ChatGPT about all of my data, and it has concluded that I am what's called "A Lean Mass Hyper-Responder (LMHR)" - **someone, typically lean and metabolically healthy, who experiences a significant rise in LDL cholesterol (high LDL) when they drastically cut carbs and increase fat intake (like on a keto diet), while also having low triglycerides (TG) and high HDL cholesterol (high HDL)**. This combination (high LDL, high HDL, low TG) is called the "[Triad](https://www.google.com/search?q=Triad)," and while it signals excellent metabolic health markers, the very high LDL levels concern some, though research is ongoing about the long-term cardiovascular risk.

The problem here is that Function has path for identifying this, so it just keeps telling me I have high cholesterol and suggests that I stop eating processed foods and foods high in fat. But I eat an insanely healthy diet, I'm the same weight that I was in 9th grade, I workout 5-6 days a week, there's not much more I can change, but Function can't seem to come to the same conclusion ChatGPT can.

This is where I think connecting all this data together, and running an LLM like ChatGPT on the backend could be really interesting. What ChatGPT did that Function didn't is asked for my Triglyceride and ApoB levels, once I gave it that data, in combination with my cholesterol it said:

> **This strongly supports the lean-mass hyper-responder phenotype**

> LMHR is distinguishable from classic high-LDL patterns by its *lack* of metabolic dysfunction markers.

> Your numbers are practically the textbook:

> • ApoB normal
• Triglycerides low
• hs-CRP near zero
• Lean, physically active, daily fasting
• LDL-P and small LDL somewhat elevated

> This is the LMHR “fingerprint.”

But Function just keeps telling me to stop eating fast food and drinking sodas...but I haven't had either in years. If I could connect all of this data together, like I'm doing manually with ChatGPT, then I can actually get the insights I need.

As you can tell, I've been doing this manually, hoping that someone will come out with something that connects all of this data together. But it doesn't seem to be happening, and I'm getting sick of talking to a Chatbot, this can, and should be done agentically.

Maybe someone has already built this? If not, I might just have to do it myself! Okay, this ended up a lot longer than I expected, so I'll end it here.

Not written by AI so I apologize for any typos or bad grammar, just wanted to get my thoughts out here!
