Live Transcribe unresponsive

Using /e/OS Applications
1

Hello

I am deaf (and with constant tinnitus, hence my username), and I have recently discovered Live Transcribe, which works with my mother’s Android, but I understand it doesn’t work with de-googled Android. Indeed, I can install and open the app on my own phone, but it just doesn’t type out anything I say.

I’ve read the only other thread I found here about it, but as I know nothing about the workings of computers or phones, I have no idea about the meaning of what I read.

Can anybody help at all? In baby language?

Thank you.

Regain your privacy! Adopt /e/OS the deGoogled mobile OS and online services

phone
phone427×374 173 KB

2

Replacement for Live Transcribe app?

3

the other relevant thread is: Captioning phone calls for hard-of-hearing users

is that your exact use-case, live transcription of incoming calls?
or generally speech to text when the other person can speak into your microphone?

4

Hi again piero

Yes, that’s the thread I already looked at and didn’t make much sense of. I will see if there are any alternatives, but so far the similar apps I have looked at are not specifically for deaf people needing to see what people are saying. They are more focused on memos and other features. Live Transcribe is simple and does just that one important job. But I will keep looking.

5

Hi tcecyk

Thanks for the link to that other thread, but no, it’s not what I am searching for. I already use Relay UK for calls, which is fine (and the reason I chose /e/foundation OS, because it at least included the Relay UK app). But thank you anyway.

6

Did you notice the recommendation in the thread, “Ava”?

Murena worked to integrate transcription for subscription users, it will use openai as backend. Proxied, but still biometric. Could be an backlog issue to offer it in an extra App, outside the input keyboard mic symbol.

I looked at where the Google live transcribe app fails (log). One needs to add the package hash here to potentially overcome that APP_CERT permission error cite in the log. I can follow up on this, but it takes a bit of time to trickle back and I can imagine it’s not the only error.

7

I’m unsure what privilege is needed for APP_CERT so I filed an issue instead of a pull request

8

Hi tcecyk

Apologies! I replied, but neglected to hit the ‘reply’ button. Or that’s how it looks now I have come back to this website, and see the box still here with my message in it.

What I wrote that appears to be pending, is:

“No, I didn’t notice Ava. I didn’t read the whole thread as I thought it wouldn’t be applicable, as the OP was asking about phones. I now have Ava on my phone, and after a quick try it seems to work just as well as Live Transcribe. So thank you very much for alerting me to this.”

Strange thing about Ava is that it makes my phone get hot pretty quickly. But anyway, I now have something on my phone that works okay (I chose the free option, which means there are more errors with it, but it’s better than nothing).

Thanks for making your request for Live Transcribe too.

1 Like
9

high battery consumption can be found in Avas app reviews. It has an offline mode, as to online mode I’d think it wouldn’t incur such high a load.

I still think its worthwhile to register the app type need (split-window transcription outside general text-input, basically just clone that Trascribe app) with the backlog. If you have a murena-ID you can use that to register in gitlab.

Murenas voice-to-text is not ideal with using openai as remote transcription, but isn’t much different than using other services. Imo 1-2 years and they can host a good model themselves or even do it on-device.