[Looking for software] voice recognition/dictation application that is FOSS (pawb.social)

submitted 3 weeks ago by VinesNFluff@pawb.social to c/foss@beehaw.org

11 comments fedilink hide all child comments

Title.

Don't really want to give my soul to a tech megacorps, but I do a lot of writing for work and hobbies, and sometimes carpal tunnel gets the best of me.

top 11 comments

sorted by: hot top controversial new old

[-] als@lemmy.blahaj.zone 7 points 3 weeks ago

I don't know what platform you're using but for desktop linux I use Nerd Dictation and for android I use FUTO Keyboard, both are completely offline and FOSS.

[-] VinesNFluff@pawb.social 4 points 3 weeks ago* (last edited 3 weeks ago)

Ah sorry. Linux Desktop, yes.

EDIT: I have now tested Nerd Dictation for a bit.

I find it funny how the English model doesn't get a single word wrong, even through my 12 inch thick brazilian accent

.... But the portuguese-brazilian model could barely understand anything I was saying.

More an observation than anything. It amused me.

[-] halm@leminal.space 4 points 3 weeks ago

Ah yes, the anglocentrism of software development. I gave up on Google translate years ago because it clearly only catered to English speakers. The other way around was such ham fisted mistranslations with English syntax, it was basically useless.

[-] badcodecat@lemux.minnix.dev 1 points 3 weeks ago

i believe FUTO keyboard isn't FOSS

[-] VinesNFluff@pawb.social 3 points 3 weeks ago

You are right-ish

They have a source-available license, which allows all modification and redistribution except modification and redistribution with commercial purposes.

How you feel about that is up to you. Personally I feel like allowing commercial usage of the code is actually a major blindspot with the normal FOSS licenses, as it leaves the doors open for TiVoization.

[-] Vitaly@feddit.uk 3 points 3 weeks ago

https://github.com/openai/whisper

[-] IrritableOcelot@beehaw.org 5 points 3 weeks ago

Uhhh is that not pretty much the definition of a "tech megacorp"?

[-] VinesNFluff@pawb.social 2 points 3 weeks ago

Teeeeeeeeeechnically what they linked is for running the AI model inside your own machine, and so not much is given to OpenAI when you do so other than I guess attention.

.... But yes, they are absolutely a tech megacorp. It then depends on how you define "giving your soul" when it pertains to said tech megacorp.

[-] CaptainAmeristan@lemmy.zip 3 points 3 weeks ago

What about Speech Note? I have used it for transcription of Zoom calls and it does a pretty good job! GitHub link

[-] VinesNFluff@pawb.social 2 points 3 weeks ago

Gave it a shot too (it even has an option to use the same language model as nerd-dictation cited by another poster)

It works. Once again, works very well in english and just gives up in portuguese.

I guess I'll just have to accept that computers are anglo-only tools. [cackles]

[-] halm@leminal.space 2 points 3 weeks ago

I've been looking at Vosk, but haven't found a good client for their API yet. Anyone else?

this post was submitted on 30 Oct 2024

19 points (100.0% liked)

Free and Open Source Software

17949 readers

84 users here now

If it's free and open source and it's also software, it can be discussed here. Subcommunity of Technology.

This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.

founded 2 years ago

MODERATORS

Gaywallet@beehaw.org

alyaza@beehaw.org