Speech-to-Text app with Offline PC
Speech-to-Text app with Offline PC
I am looking for recommendations for a decent Speech-to-Text application (Whisper) which I can download & install on an MX Linux machine without any access to internet.
A self-contained appimage or an appimage with a separate Speech Model download (that I can transfer to the offline PC and point it to) would be as complex as I'd want to make it.
kdenlive requires internet access for its speech-to-text stuff. All the offerings that I have looked at require rather difficult installation on internet-enabled computers. I want an app that runs as portably as possible, as simply as possible.
A self-contained appimage or an appimage with a separate Speech Model download (that I can transfer to the offline PC and point it to) would be as complex as I'd want to make it.
kdenlive requires internet access for its speech-to-text stuff. All the offerings that I have looked at require rather difficult installation on internet-enabled computers. I want an app that runs as portably as possible, as simply as possible.
Re: Speech-to-Text app with Offline PC
Try Speechnote, Flatpak
Desktop Mobo: BIOSTAR model: B450MHP
CPU AMD Ryzen 3 3200G
RAM: 8 GiB 2667 MHz DDR4
Device-1: AMD Picasso/Raven 2
STORAGE:Gigabyte model: GP-GSTFS31480GNTD size: 447.13 GiB
DISTRO: MX-23.6_x64 Libretto XFCE 4.20
CPU AMD Ryzen 3 3200G
RAM: 8 GiB 2667 MHz DDR4
Device-1: AMD Picasso/Raven 2
STORAGE:Gigabyte model: GP-GSTFS31480GNTD size: 447.13 GiB
DISTRO: MX-23.6_x64 Libretto XFCE 4.20
Re: Speech-to-Text app with Offline PC
I might have this wrong, but doesn't downloading a flatpak version of anything only allow install onto the PC I downloaded it from, and not allow me to run it on a completely offline PC?
Re: Speech-to-Text app with Offline PC
https://github.com/Const-me/Whisper
So this is the type of thing I am looking for, but for MX Linux.
I can download the app & the larger speech-to-text Model file. I transfer them both to a completely offline machine that never sees the internet, running it portably.
I suppose it's possible for me to run this in a Windows virtual machine, but I'd rather not run in a virtual machine if I can avoid it. The (relative) slowness, the extra RAM usage, etc.
So this is the type of thing I am looking for, but for MX Linux.
I can download the app & the larger speech-to-text Model file. I transfer them both to a completely offline machine that never sees the internet, running it portably.
I suppose it's possible for me to run this in a Windows virtual machine, but I'd rather not run in a virtual machine if I can avoid it. The (relative) slowness, the extra RAM usage, etc.
- DukeComposed
- Posts: 1395
- Joined: Thu Mar 16, 2023 1:57 pm
Re: Speech-to-Text app with Offline PC
The literal first line of the README states "This project is a Windows port of the whisper.cpp implementation."user101 wrote: Mon Jun 02, 2025 3:13 am https://github.com/Const-me/Whisper
So this is the type of thing I am looking for, but for MX Linux.
I suppose it's possible for me to run this in a Windows virtual machine, but I'd rather not
-
- Posts: 38
- Joined: Thu Feb 04, 2021 11:57 am
Re: Speech-to-Text app with Offline PC
You might want to take a look at the “Hugging Face” solution, which is used in the June project ( https://github.com/mezbaul-h/june ). Voice input / processing (AI) / voice output locally on the (MX) computer WITHOUT internet access.
I would be very happy if this would help you.
I would be very happy if this would help you.
2025: Es gibt keine schlechten Linux-Distributionen mehr.
2025: There are no more bad Linux distributions.
2025 : il n'y a plus de mauvaises distributions Linux.
2025: There are no more bad Linux distributions.
2025 : il n'y a plus de mauvaises distributions Linux.
-
- Posts: 301
- Joined: Fri Jul 17, 2020 10:32 am
Re: Speech-to-Text app with Offline PC
Two assumptions reading between the lines :), 1) you mention kdenlive's speech-to-text stuff, so I assume speech-to-text processing of material within media production software, for transcribing/subtitling purposes or similar, would be the main goal (instead of, say, realtime dictating of spoken shell commands), and 2) if the first assumption is correct, then you are generally at least somewhat familiar with media production software.user101 wrote: Fri May 30, 2025 11:20 pm A self-contained appimage or an appimage with a separate Speech Model download (that I can transfer to the offline PC and point it to) would be as complex as I'd want to make it.
kdenlive requires internet access for its speech-to-text stuff. All the offerings that I have looked at require rather difficult installation on internet-enabled computers. I want an app that runs as portably as possible, as simply as possible.
If these assumptions match, then: https://techaud.io/projects/reaspeech-lite/ ... Check that out even if Reaper isn't already familiar to you beforehand. This is the most streamlined free and offline solution for media transcription purposes I currently know of. It started out as a Docker solution combined with a Reaper Lua script, you can see some of how it came about (and what to expect) in this thread https://forum.cockos.com/showthread.php?t=294811 - but that's more like, for background and informational purposes, as this more recent plugin based solution is more straightforward to use. Especially if production software like this is already familiar to you, that is.
Re: Speech-to-Text app with Offline PC
If you can work out how to download whisper.cpp, then transfer the files onto a completely offline PC that will never see internet, let me know.DukeComposed wrote: Mon Jun 02, 2025 3:53 am
The literal first line of the README states "This project is a Windows port of the whisper.cpp implementation."
Re: Speech-to-Text app with Offline PC
I am aware of many apps that work without internet access (after they've been installed using the internet). It's just that I want the app to be able to be totally portable and self-contained: download it on one internet-enabled machine, then transfer files to completely offline machine & have it run successfully.KaramellPudding wrote: Mon Jun 02, 2025 5:02 am You might want to take a look at the “Hugging Face” solution, which is used in the June project ( https://github.com/mezbaul-h/june ). Voice input / processing (AI) / voice output locally on the (MX) computer WITHOUT internet access.
I would be very happy if this would help you.
Re: Speech-to-Text app with Offline PC
Follow the Quick Start in the Readme, it's really not that hard to compile and move it wherever you want.user101 wrote: Mon Jun 02, 2025 11:50 pmIf you can work out how to download whisper.cpp, then transfer the files onto a completely offline PC that will never see internet, let me know.DukeComposed wrote: Mon Jun 02, 2025 3:53 am
The literal first line of the README states "This project is a Windows port of the whisper.cpp implementation."
https://github.com/ggml-org/whisper.cpp
Where's the problem?