Confused how to run on MacOS

[divan]

Hi, fantastic project! I followed installation steps for MacOS, all went smoothly, but now stuck with two things:

1. Hotkeys

There is no 'default' hotkeys for MacOS in config.py comments section.

Defining hotkeys is often confusing topic in other apps (is it cmd or CMD, 'super'-key etc), but then here is a line in config.py that says:

# On Mac/Linux, a hotkey cannot overlap another (e.g. cmd+e and cmd+shift+e)

What does it mean? Should I check system hotkeys and make sure that chosen hotkey is not overlapping?

I added 'cmd-shift-m', but nothing happens. I don't know if something is wrong with hotkey or it's another issue. (That's why it would be cool to have default hotkeys)

2. I see a line in the output, which probably explains why hotkeys is not working.

This process is not trusted! Input event monitoring will not be possible until it is added to accessibility clients.

The concept of adding apps to the trusted list is not unusual on MacOS, but quick search on Readme reveals nothing about it. How do I do it? Should go to settings and add that bash script run_AlwaysReddy.sh to the list? Or it's something else completely?

Bottom line, it's a bit frustrating how smoothly the installation went, and how it's impossible to use the app out of the box because of the main UI promise (just hotkeys) needs some non-documented configuration first.

[divan]

Anyone using this project on MacOS could share their configuration and setup perhaps?

[divan]

Figured out second issue - need to add terminal to the Accessibility list in Settings->Privacy (in my case it's 'Warp' terminal)

Trying different hotkeys now, still can't make it work (

[divan]

Next, we need to add terminal to the Input Monitoring too in Privacy->Settings:

That finally makes hotkeys work. My current setup (just replaced ctrl with cmd):

CANCEL_HOTKEY = 'cmd+shift+e'
NEW_CHAT_HOTKEY = 'cmd+shift+w'
RECORD_HOTKEY = 'cmd+shift+r' # Press to start, press again to stop, or hold and release. Double tap to include clipboard
READ_FROM_CLIPBOARD = "cmd+shift+c"
TRANSCRIBE_RECORDING = "cmd+shift+t"

[ghNico]

@divan can you run AlwaysReddy with this setup? I wanted to try this out, but keep getting stuck after i told anything.

My Configuration is local lm studio and it worked only one time but without my told-text

It seems like its getting stuck during the fasterWhisper run, when i try to install via setup.py the 2nd option TransformerWhisper it installs nothing. I think the project is a bit broken. Would love to hear back from you.

TTS works, cmd+shift+c works really well, but not the STT part into LLM

[scalar27]

I tried AlwaysReddy several months ago and ran into problems (I have a M1 Mac) but then found https://github.com/DissonanceTK/MacReddy which is a fork. It works great, but has not been updated in a while. So at this point I may be missing new features or functionality. Regarding hotkeys, changed one line in config.py to RECORD_HOTKEY = 'alt+cmd' since that is easier to type. Hope this helps a bit.