I hacked together an Android app specifically for the purpose, so it doesn't really answer the original question about generally available projects. (I haven't got round to releasing the code as it's rather bespoke, but could do if you're interested.)

I'd be interested in the general technique, "teach a man to listen for beeps..."

I FFT incoming sound buffers, and if the 2KHz band is the highest for a suitable number of consecutive frames, then I send the email and stop listening for a few minutes to avoid double notifications.

2KHz is the frequency of my machine's beep. Just triggering on overall noise level alone would notify me each time the machine goes into spin, but using the FFT allows it to be more discerning.

(I'll try to publish the code soon.)

I'd appreciate seeing your code for this - I've intended to write a similar app for listening to an old alarm system I have.

I've now put it on GitHub: https://github.com/martin-stone/AndroidAudioNotifier

Thank you!

Thanks for sharing the code.

Perhaps someone (with more patience than I for Android/Kotlin development) can tweak it so that is can listen for knocking on a door to trigger some sort of alert (email/txt/broadcast an intent).

Triggering an intent would be very helpful as then it could then be acted upon by Tasker or Automagic4Android.

this would be very interesting!

Code is now on GitHub here: https://github.com/martin-stone/AndroidAudioNotifier

If you're looking for an off the shelf solution, I turned an old Android phone into a baby monitor for keeping track of crying when I'm on the other side of the house. There's plenty of baby monitoring apps that do anomalous noise detection and raises an alert.