Closed carstenschwede closed 1 year ago
Solved by running inside of Terminal instead of iTerm, likely because of missing access permissions.
It appears that the permission request does not bubble up.
I have the same problem. In iTerm I get [1] 2589 abort hear
, even with sudo
. Though it works i Terminal.app
.
I tried resetting microphone permissions using tccutil reset Microphone
, but to no effect.
Here's the output of sudo dtruss hear -m
if that could help diagnose the problem:
I just tried granting iTerm microphone permissions by running ffmpeg -f avfoundation -i ":1" -t 10 test.mp3
and allowing. Now iTerm shows up in Security & Privacy under Microphone, but hear
gives me the same output as before.
I don't get it.
@sveinbjornt Any idea what the problem might be?
You'll need "Microphone" and "Speech Recognition" permissions for hear
to work.
You might try whisper
instead which will run speech recognition offline (and depending on the model used, will outperform Apple's service). There is a repo here which provides "real-time" transcripts via microphone.
This is still an issue.
This has to do with access permissions, and has nothing to do with "hear" per se.
the main issue is the terminal requires both microphone and speech recognition permissions. there is my solution:
sqlite3 ~/Library/Application\ Support/com.apple.TCC/TCC.db
insert into access (service,client,client_type, auth_value, auth_reason,auth_version) values ('kTCCServiceMicrophone','com.googlecode.iterm2', 0, 2, 2, 1);
insert into access (service,client,client_type, auth_value, auth_reason,auth_version) values ('kTCCServiceSpeechRecognition','com.googlecode.iterm2', 0, 2, 2, 1);
.quit