Introducing Tunum: iOS app that continuously analyzes your health data with personalized AI

Hi folks,

Fellow health optimizer & quantified self enthusiast here!

I wanted to share a quantified self iOS app that I’ve been building over the last several months. Tunum securely connects & continuously analyzes your wearable, lifestyle & biomarker data to help you optimize your health! Here are some highlights:

  • Privacy-centric – your health data is securely stored only on your phone
  • Data integrations – connect popular wearables (Apple Watch, Garmin, Oura) and sync lifestyle data (caffeine, nutrition, meditations and more) from hundreds of apps via Apple Health
  • Scan documents – automatically extract and consolidate biomarker data (blood test results and body composition scans) – from images & PDFs
  • AI-enhanced – get actionable daily summaries & review your health data with AI
  • Advanced analysis – uncover trends & correlations and understand lifestyle impact on sleep, recovery and other health metrics
  • Daily journal – log your caffeine, mood, sleep habits and more
  • Data export – export your health data as a CSV file with just a single tap

We’re building this app primarily for Quantified Self members like you, so would really appreciate any feedback! If the app is missing any features you would like to see, please comment or DM me.

If you’d like to check it out, there are both free & pro versions of the app. Pro version comes with a free 2-week trial – DM me for a special discount for Quantified Self members if you’re interested!

Thank you!

P.S. We’re actively working on adding more features, with workouts and CGM support next on our list!

2 Likes

This sounds amazing! Congrats on getting something so ambitious to launch.

Sadly I’m on android so I can’t test this out but I will try to find an iPhone to give it a go.

I’d like to build something that has some overlapping features to yours - specifically I’ve been researching Garmin and Fitbit data integrations. My research landed me on an application process for developer API access and this asked for things including website, app user numbers etc. It seemed to me that they expected the API to be used in only established apps, and pre-launched apps might struggle to get permission.

Did you have any problems like this?

Thanks in advance

1 Like

Can you say something about how the the data is stored and how the export function works? I looked at your web site for documentation but didn’t see any beyond general terms of service warnings that the data is fully under your control and no promises are made about storing, deleting, or notifying anybody about changes or withdrawal. Almost everybody who has participated in this forum for any length of time has had the sad experience of losing data entrusted to aggregators.

1 Like

Thank you! We’re hoping to add Android support too at some point, as we’ve heard that feedback from interested users

Yes, this is a pretty standard process re: getting developer API access – in my experience you do need to have a website and ideally an incorporated company. They sometimes ask about app user numbers but I believe that it’s only to address any scaling concerns for their backends. Most providers also have a simpler process to get development access (usually with a limited number of users that can use your integration) and an additional review to get production access with no limits.

Hope that helps!

Great question! Also been there myself with losing data and heard a lot of horror stories too

The data is currently stored in a local database on your phone, we don’t see it or control how it’s stored or deleted. We’re also planning to add an encrypted iCloud backup to prevent people from losing their data. The export function is very simple: you pick a date range, and we create a CSV file with all the health records that you have in the local database on the phone (both synced from the wearable integrations, as well as the manual data you put in such as journal logging, bloodwork, body scans, etc). You can save that file in the Files app on iPhone or share it with anyone you like (e.g. email, message, etc)

Hope this helps!

1 Like

Super helpful and encouraging.

Thanks and good luck

1 Like

Could you tell me more about this? Will I see change point, shape or cycles detection?

1 Like

Here is a high level overview:

  • trend analysis – you can view simple week-over-week, month-over-month trends (i.e. % changes) + get the biggest up/down movers automatically surfaced for you daily. we’re considering adding change point detection too

  • correlation analysis – you can correlate (Pearson correlation) last 3 months of data across all series that have daily values + get the significant positive/negative correlations automatically surfaced for you daily. you also can add 1-day lag when correlating series

  • event analysis – you can create custom events based on any numerical or yes/no time series. for example, an event could be “awake duration (during the night) > 1 hour”, “had 2 cups of coffee after 9pm”, “had 500 mcg of melatonin”. then you can view difference in the averages for the dependent metric (e.g. “deep sleep”, “recovery level”, etc.) on event days vs other days. for example, “how is my deep sleep impacted by 2 cups of caffeine after 2pm the day before”. additionally, you can create custom comparisons, e.g. “2 cups of coffee after 2pm the day before vs 1 cup of coffee after 2pm”

hope this helps! curious if there are also other types of analysis that you/others would find useful

1 Like

Hello,

Seems like a nice App.

Little out of context,
please suggest some free open source Integrated Universal API to access users’ health data from various wearables, RPM devices, and mHealth apps like Metriport.

Integrated API should support different integrations such as:

  • Apple Health
  • Google Health
  • Fitbit
  • Oura
  • Whoop
  • Withings
  • Cronometer

Thanks.

Unfortunately, I am not aware of any free ones. The only ones I know about are Metriport and Terra (https://tryterra.co), but neither is free or open-source

I recently found Rook https://docs.tryrook.io/
Also not free, but potentially cheaper than Terra.

There are many such tools. Look them up on the open humans wiki.

1 Like

Most providers sync into Apple Health and Google Fit, so you could probably sync other providers by proxy to meet your needs. That’s the strategy of the product our nonprofit built, at least.

Whoop, Oura, and Withing, sync data to Apple Health that can be synced/exported elsewhere.

This is true, but it also depends on the use case and how granular you want the synced data to be. For example, Apple Health won’t have information about sleep stages when you sync data from Whoop/Oura. And some data is only available via API and is never written to Apple Health

That’s odd. I’m seeing Oura sync sleep stages to Apple Health. I agree with the concern about some data only being available via API, though. On top of that, sometimes Apple data exports don’t even include all your data (like medication dosages).

What AI engine will this app use? Many of the most prominent ones extract the data and stick it back into their LLM-database, which would militate against the privacy-centric ideal should one of those be being used.

1 Like

Good question! We use OpenAI API at the moment, but they actually do not train their LLMs on the data from APIs and don’t even retain that data. The only customer data they train on is the conversations from ChatGPT. I think long term ideal though is LLMs that are running on your own device without any 3rd parties :slight_smile: Hoping to get there soon, but the technology is still not fully there

ah, you’re totally right about Oura, looks like they started syncing sleep stages in the fall 2023, and I was doing research on the integrations prior to that. it used to be just “Asleep”/“Awake”/“In bed”, which seems to still be the case for Whoop too.

and agreed, Apple data export is super inconsistent

Suppose I want to access data from various health apps and sources and store it to some local database then how should I proceed to achieve this using metriport api or any other same api like this? Any guide would be helpful.

Hi, the app looks interesting. Any plans to make it available outside the US?