Passive sensing is the way to go
Foundational Health reads sound levels from your microphone, but never records audio. Here's how that works, why we made that choice, and what it means for your privacy.
One of the first decisions we made building Foundational Health was this: we will never record or store audio.
It sounds like a constraint. In practice, it’s the thing that makes the product possible. Here’s why.
What we actually need from sound
For the purposes of understanding your environment, we don’t need to know what you’re hearing. We need to know how loud it is, and roughly what kind of sound it is, e.g steady, intermittent, sudden.
How we do it
On modern phones, we can compute these features on the device, in real time, and discard the raw audio immediately. The microphone is read for a fraction of a second, the level is calculated, and the audio buffer is overwritten. The only thing that persists is a number.
This has two big advantages:
- Privacy. Your conversations, your music, your voice never leave the microphone buffer. There’s no recording to leak, to hack, to subpoena, or to accidentally upload.
- Battery. Processing a few samples per second is vastly cheaper than continuously recording and streaming audio.
Why we’re loud about this
A lot of apps collect more than they need because the data is valuable to advertisers, to data brokers, to whoever. We’re building Foundational Health for people whose sensory data is especially intimate, and for whom trust has been broken before. So we made the choice to collect less, and to be explicit about it.
That means:
- No raw audio, ever. Not on the device, not on our servers.
- No transcripts, no speech-to-text, no “just this one feature.”
- If we ever can’t compute something on-device, we’ll tell you, and it’ll be opt-in.
What this means for you
When you see the app reporting “42 dB, calm” in your timeline, that’s the only thing that was ever measured. There is no recording behind it. If you delete your data, it’s gone.
We think that’s the right tradeoff. You get the insight “your office is loud and you feel worse there”, without the surveillance. If that’s the kind of app you want to exist, join the waitlist and help us build it.