On a functional level itâs the fastest & the most convenient input method when youâre not in front of the proper keyboard. Which even for me constitutes a big chunk of my life đ.
Iâve recently made a few updates to my voice notes flow, removing various sources of friction and streamlining the experience. I figured Iâd share a few details about that so other people can replicate it.
New Flow
Use a smartwatch as the main input device
Itâs configured in a way that recording is started on double-tap of one of the watch buttons
Two issues with the older iteration of the process motivated this bout of improvements
when I go to sleepâI leave my phone (which used to be the main input device for taking voice notes) away from my bed to avoid doomscrolling instead of sleeping sleep hygiene
but often when I drift away to sleep (or when I wake up at 3am), Iâd have a thought that Iâd want to record. And as I wonât have a phone on me, I wonât be able to do that.
Iâve tried screaming at Alexa or my phone from across the room to take notes, but that never worked well. Neither was I happy with the approach of having a physical notebook close to my bed (lighting questions + it doesnât really fit into my general knowledge management practices).
even when I had the phone nearby, the experience of taking a note involved an unfortunate amount of friction
reach for the phone (which can be surprisingly non-trivial, e.g. when biking, or just getting it from tight pants lol)
unlock it (ok, technically I found a way to start recording from lock screen, but it involved similar amount of work to unlocking the phone, so đ¤ˇââď¸)
These issues suggested a shape of the solutionâI needed a dedicated device that would allow me to start recording with minimal friction, and wonât be a distraction in a way the phone could be.
Originally this made me look for dedicated voice recorders with ability to automatically sync recordings online
Surprisingly I wasnât able to easily find a device that would fit the requirements here.
Iâve considered adopting Go Note Goâa project a friend of mine created for similar purposes. But I didnât really want to deal with custom hardware.
Eventually Iâve settled down on the smartwatch as the form factor
it is arguably more convenient compared to a necklace format
and it is a more âsocially acceptedâ tech thing to wear
I also feel like it exists in a nice middle-ground between a single purpose device and general-purpose phone. Allowing me to derive additional utility from displaying information or programming it to trigger automations, while not unnecessarily consuming my attention.
One thing I was not certain about with regard to smartwatch form factor is whether Iâd be able to get the friction to starting the recording low enough. As mobile devices tend to be unnecessarily constrained in how much users can customize them.
Amazon return policies and an inkling that the baseline experience is not too bad led me to give it a shot.
It took some work, but I was able to get to a place where it takes just one action (double-click) to start the recording, which Iâm pretty happy with.
The rest of the post describes technical details on how to get there, as I dearly wish someone would have written this guide before.
How to get Android Wear watch to start voice recording in one click
Tools
Galaxy Watch 6
This is the watch I have. Motivation for this specific model is that itâs the latest android watch model I could get for cheap đ. And 80% of my intended use is just voice notes đ¤ˇââď¸
Another motivation was that Iâve read online that Samsung watches built-in voice recorder is pretty good
To further streamline the experience we need to dive into the world of Android Macrosâspecifically, weâd be using AutoWear
Autowear is a very confusing app to use, and mostly not designed to be used on itâs own (the idea that youâd use it with Tasker, Macrodroid or similar)
But it allows you to do things not possible at baseline, in particular weâd be using itâs ability to click elements on the screen to automatically start a recording.
Steps
Install AutoWear on both phone and watch
Start the AutoWear trial on the phone
if you donâtâit will just give you confusing errors when you try to configure actions (without telling you that the lack of subscription is the reason)
or you can purchase it (1.6$), but trial is sufficient to do initial setup and things will continue working after it expires, so you technically donât have to buy it
but after trial expires updating configuration will not work (with the aforementioned confusing errors)
Enable one of the Launcher Apps to serve as a command trigger
On the watch, go to AutoWear â Launcher Apps
Enable one of the apps there (I use âCommandâ),
This would make an icon for it appear in the list of watch apps
And youâd be able to map your double press button to this âAppâ
Enable accessibility access for AutoWear on the watch to make âinputâ commands work *
AutoWear on the phone
In the app on the phone weâd need to configure two actions to run when the launcher app is opened
App Action (to actually open the voice recorder app)
Launch App Package
com.coffeebeanventures.easyvoicerecorder
Advanced
Command To execute
Select âApp open Commandâ option
Results in &APPOPENEDCOMMAND& text value for the command
Input Action (to click start recording button)
Command To execute
Select âApp open Commandâ option
Results in &APPOPENEDCOMMAND& text value for the command
Command
Help setting up command â yes
(Open the Easy Voice Recorder on the watch & make sure the watch screen is on when youâre configuring this)
âClick Elementâ
Select âtextâRecordâ as item to click
Results in click(text,Record) text value forthe command
Technically these two are configured as independent actions that both happen when you open the launcher app (instead of being a proper workflow where steps happen in order)
This works because the Input action has a 5 second timeout and within those 5 seconds it keeps watching for relevant element to appear on the screen and it clicks it whenever that happens.
This allows us to set things up with AutoWear only, without relying on Tasker/âMacrodroid/âetc and needing phone-watch communication for this to work
It takes 1.5-3 sec for recording to start after you press the button
Which is forever in computer terms. I think part of the reason for this is a cobbled together automation flow nature of this, but also just launching the app is not really instantaneous =\.
Stopping the recording
To stop the recording, I need to wake up the screen and tap a button on it.
Ideally Iâd just double-tap the trigger button again to stop the recording.
This is probably something I can do with Macrodroid/âTasker, but havenât looked into it yet and suspect it wonât be great for performance reasons.
When you grant accessibility access to any app, the watch starts making a click sound when screen locks and you canât turn this off đ
I found this very frustrating (though I suspect I might be relatively more noise sensetive then avg).
Thankfully, there is a mitigationâyou can just lower the Media volume on the watch.
This works well for me bc I rarely if ever want to play sounds on my watch. Also, there is a sweet spot, where I can still hear e.g. the Google Assistant replies, but the click becomes imperceptible.
Misc tips
By default, the watch will hide the last app and show you the home screen, after 20s of screen turning off.
Which I found annoying for voice notes use case bc, that added number of clicks I needed to do to stop the recording.
You can make this better by changing the Display > Show last app setting to a larger value (pretty happy with 2 minutes)
Native Samsung Voice Recorder also very nice, and is what I originally used, but Easy Voice recorder has a slightly better UX for my purposes.
Specifically when recording stopsâSVR navigates to ârecording listâ app screen. Which I never really want because I donât want to listen to my recordings on the watch, and especially not after I just made one.
When the app is in that stateâthe flow canât be triggered again, until you navigate âbackâ to main screen.
So stopping the recording is actually 2 actionsâtap the stop button and swipe back, which is a bit annoying. In part because Iâd neglect to swipe back immediately a lot of the time, and then would have to do it next time I trigger the flow & that would ruin the fluency of starting it in just one action.
EVR, on the other hand, gets into the state to start a new recording almost immediately (after a brief confirmation screen).
EVR also has native cloud upload capabilities, which can be convenient for setting up further notes processing similar to what I do in Voice notes inbox process
SVR has native transcription capabilities, but they are not as good as using Whisper or similar (which is what I do)
Finally, EVR is developed by a real human that I can talk to đ
Which gave me hope that maybe I can request changes/âtweaks to further optimize this flow.
That has been a mixed success so farâthe support replied to my messages, but itâs unclear to which degree is the app in active development, so I donât have a strong expectation for changes I requested making it through in the near future.
Streamlining my voice note process
Link post
Why voice notes?
On a functional level itâs the fastest & the most convenient input method when youâre not in front of the proper keyboard. Which even for me constitutes a big chunk of my life đ.
Taking a voice note allows me to quickly close an open loop of a stray though or a new idea and go on with my lifeânow mentally unburdened
Another important aspect is that taking voice notes evokes a certain experience of âfluencyâ for me.
That is because when youâre taking a voice note, you can record a stream of consciousness without having to re-formulate or filter things.
Which allows you to avoid switching away from the context of original thought/âbreaking the flow to enter an âeditor modeâ.
This is the reason that Iâd occasionally take voice notes even if Iâm in front of the proper keyboard.
See Iterating fast: voice dictation as a form of babble for more on this.
Itâs worth noting that Iâm specifically talking about taking a voice note vs doing voice typing/âdictation in this context.
Voice notes allow me to better harness the fruits of diffuse thinking mode (shower thoughts)
Iâve recently made a few updates to my voice notes flow, removing various sources of friction and streamlining the experience. I figured Iâd share a few details about that so other people can replicate it.
New Flow
Use a smartwatch as the main input device
Itâs configured in a way that recording is started on double-tap of one of the watch buttons
Notes are then automatically uploaded to the cloud, transcribed and added to my main knowledge management system
Two issues with the older iteration of the process motivated this bout of improvements
when I go to sleepâI leave my phone (which used to be the main input device for taking voice notes) away from my bed to avoid doomscrolling instead of sleeping sleep hygiene
but often when I drift away to sleep (or when I wake up at 3am), Iâd have a thought that Iâd want to record. And as I wonât have a phone on me, I wonât be able to do that.
Iâve tried screaming at Alexa or my phone from across the room to take notes, but that never worked well. Neither was I happy with the approach of having a physical notebook close to my bed (lighting questions + it doesnât really fit into my general knowledge management practices).
even when I had the phone nearby, the experience of taking a note involved an unfortunate amount of friction
reach for the phone (which can be surprisingly non-trivial, e.g. when biking, or just getting it from tight pants lol)
unlock it (ok, technically I found a way to start recording from lock screen, but it involved similar amount of work to unlocking the phone, so đ¤ˇââď¸)
click a button to start recording
put the phone away
These issues suggested a shape of the solutionâI needed a dedicated device that would allow me to start recording with minimal friction, and wonât be a distraction in a way the phone could be.
Originally this made me look for dedicated voice recorders with ability to automatically sync recordings online
I was thinking something in form factor similar to https://ââwww.friend.com
Surprisingly I wasnât able to easily find a device that would fit the requirements here.
Iâve considered adopting Go Note Goâa project a friend of mine created for similar purposes. But I didnât really want to deal with custom hardware.
Eventually Iâve settled down on the smartwatch as the form factor
it is arguably more convenient compared to a necklace format
and it is a more âsocially acceptedâ tech thing to wear
I also feel like it exists in a nice middle-ground between a single purpose device and general-purpose phone. Allowing me to derive additional utility from displaying information or programming it to trigger automations, while not unnecessarily consuming my attention.
As a bonus point, the watch being waterproof allows me to easily capture shower thoughts, which I deliberately have more space for now
One thing I was not certain about with regard to smartwatch form factor is whether Iâd be able to get the friction to starting the recording low enough. As mobile devices tend to be unnecessarily constrained in how much users can customize them.
How silly it is that I can change what one watch button does (in a very limited way), but not the other?
Amazon return policies and an inkling that the baseline experience is not too bad led me to give it a shot.
It took some work, but I was able to get to a place where it takes just one action (double-click) to start the recording, which Iâm pretty happy with.
The rest of the post describes technical details on how to get there, as I dearly wish someone would have written this guide before.
How to get Android Wear watch to start voice recording in one click
Tools
Galaxy Watch 6
This is the watch I have. Motivation for this specific model is that itâs the latest android watch model I could get for cheap đ. And 80% of my intended use is just voice notes đ¤ˇââď¸
Another motivation was that Iâve read online that Samsung watches built-in voice recorder is pretty good
Easy Voice Recorder
The voice recording app I use, there are several reasonable options but this is my current preferred one.
Automatic sync of recordings to your phone is probably the key feature you want here if youâre considering other alternatives.
Baseline experience (double-press the button + tap the screen)
This is the setup you can get to without going into the weeds (which is not terrible, but not ideal). The way to do it is to
remap the double press of home key to launch Easy Voice Recorder
then youâd need to tap the screen to the start the recording
Going beyond baseline experienceâAndroid Macros (AutoWear)
To further streamline the experience we need to dive into the world of Android Macrosâspecifically, weâd be using AutoWear
Autowear is a very confusing app to use, and mostly not designed to be used on itâs own (the idea that youâd use it with Tasker, Macrodroid or similar)
But it allows you to do things not possible at baseline, in particular weâd be using itâs ability to click elements on the screen to automatically start a recording.
Steps
Install AutoWear on both phone and watch
Start the AutoWear trial on the phone
if you donâtâit will just give you confusing errors when you try to configure actions (without telling you that the lack of subscription is the reason)
or you can purchase it (1.6$), but trial is sufficient to do initial setup and things will continue working after it expires, so you technically donât have to buy it
but after trial expires updating configuration will not work (with the aforementioned confusing errors)
Enable one of the Launcher Apps to serve as a command trigger
On the watch, go to AutoWear â Launcher Apps
Enable one of the apps there (I use âCommandâ),
This would make an icon for it appear in the list of watch apps
And youâd be able to map your double press button to this âAppâ
Enable accessibility access for AutoWear on the watch to make âinputâ commands work *
AutoWear on the phone
In the app on the phone weâd need to configure two actions to run when the launcher app is opened
App Action (to actually open the voice recorder app)
Launch App Package
com.coffeebeanventures.easyvoicerecorderAdvanced
Command To execute
Select âApp open Commandâ option
Results in
&APPOPENEDCOMMAND&text value for the commandInput Action (to click start recording button)
Command To execute
Select âApp open Commandâ option
Results in
&APPOPENEDCOMMAND&text value for the commandCommand
Help setting up command â yes
(Open the Easy Voice Recorder on the watch & make sure the watch screen is on when youâre configuring this)
âClick Elementâ
Select âtextâRecordâ as item to click
Results in
click(text,Record)text value forthe commandTechnically these two are configured as independent actions that both happen when you open the launcher app (instead of being a proper workflow where steps happen in order)
This works because the Input action has a 5 second timeout and within those 5 seconds it keeps watching for relevant element to appear on the screen and it clicks it whenever that happens.
This allows us to set things up with AutoWear only, without relying on Tasker/âMacrodroid/âetc and needing phone-watch communication for this to work
In watch settingsâremap the double press of home key to launch the âCommandâ app
Things Iâm still not happy with in the new flow
It takes 1.5-3 sec for recording to start after you press the button
Which is forever in computer terms. I think part of the reason for this is a cobbled together automation flow nature of this, but also just launching the app is not really instantaneous =\.
Stopping the recording
To stop the recording, I need to wake up the screen and tap a button on it.
Ideally Iâd just double-tap the trigger button again to stop the recording.
This is probably something I can do with Macrodroid/âTasker, but havenât looked into it yet and suspect it wonât be great for performance reasons.
When you grant accessibility access to any app, the watch starts making a click sound when screen locks and you canât turn this off đ
I found this very frustrating (though I suspect I might be relatively more noise sensetive then avg).
Thankfully, there is a mitigationâyou can just lower the
Mediavolume on the watch.This works well for me bc I rarely if ever want to play sounds on my watch. Also, there is a sweet spot, where I can still hear e.g. the Google Assistant replies, but the click becomes imperceptible.
Misc tips
By default, the watch will hide the last app and show you the home screen, after 20s of screen turning off.
Which I found annoying for voice notes use case bc, that added number of clicks I needed to do to stop the recording.
You can make this better by changing the
Display > Show last appsetting to a larger value (pretty happy with 2 minutes)Itâs worth noting that Samsungâs native voice recorder, would just stay on screen while recording is active.
If Input action stops working
This happened to me only once so far, restarting the watch fixed it.
References
Transcribing Matrix Chat voice messages with OpenAI Whisper
Voice notes inbox process
My exobrain setup
Is it possible to record audio on the watch using Autowear ? : r/âtasker
Appendix
Easy Voice Recorder (EVR) vs Samsung Voice Recorder (SVR)
Native Samsung Voice Recorder also very nice, and is what I originally used, but Easy Voice recorder has a slightly better UX for my purposes.
Specifically when recording stopsâSVR navigates to ârecording listâ app screen. Which I never really want because I donât want to listen to my recordings on the watch, and especially not after I just made one.
When the app is in that stateâthe flow canât be triggered again, until you navigate âbackâ to main screen.
So stopping the recording is actually 2 actionsâtap the stop button and swipe back, which is a bit annoying. In part because Iâd neglect to swipe back immediately a lot of the time, and then would have to do it next time I trigger the flow & that would ruin the fluency of starting it in just one action.
EVR, on the other hand, gets into the state to start a new recording almost immediately (after a brief confirmation screen).
EVR also has native cloud upload capabilities, which can be convenient for setting up further notes processing similar to what I do in Voice notes inbox process
SVR has native transcription capabilities, but they are not as good as using Whisper or similar (which is what I do)
Finally, EVR is developed by a real human that I can talk to đ
Which gave me hope that maybe I can request changes/âtweaks to further optimize this flow.
That has been a mixed success so farâthe support replied to my messages, but itâs unclear to which degree is the app in active development, so I donât have a strong expectation for changes I requested making it through in the near future.