Just thinking... Would it be possible to vibecode alternative less cluttered more accessible Crowdin interface? Not sure what their api allows...
Peter Vágner reshared this.
2026-01-23 07:59:56
2026-01-23 07:32:09
2026-01-23 07:32:09
9581444
Peter Vágner reshared this.
2026-01-23 07:31:39
2026-01-23 07:31:37
2026-01-23 07:31:37
9583674
Peter Vágner reshared this.
I just added 13 more instrumental tracks to my collection. If you need some music for your upcoming project (game, podcast, video...), Look at ondrosik.sk/music #cc0 #music
like this
Peter Vágner likes this.
Peter Vágner reshared this.
2026-01-23 07:10:12
2026-01-23 07:10:12
2026-01-23 07:10:12
9587589
Peter Vágner reshared this.
Today's crazy idea:
I've always wanted to play Elite, the old procedurally-generated spaceflight sim for the BBC Micro. It, of course, is not and will never be accessible.
However the source code is available, very small, and highly annotated. There's a whole nerdy website with bunches and bunches of details, right down to guesses on and commentaries about the relative sizes of celestial objects, ships, etc.
Tempted to hand all of that to an LLM and have it try modernizing the game in Godot, adding audio cues, speech feedback, and whatever else would be needed to make it accessible.
May not be fully faithful to the original, least of all because of 3-D audio, but I almost don't care if it captures the feel.
We'll see if I actually do this, but now the bug is lodged in my brain.
Peter Vágner reshared this.
2026-01-26 08:53:07
2026-01-23 07:07:02
2026-01-23 07:07:02
9587773
Peter Vágner reshared this.
Quite a milestone for Czech #Meshcore: Bohemia and Moravia functioned as two islands, because crossing the Bohemian-Moravian Highlands is not exactly easy. But today, the connection has been established. The two islands are beginning to function as a single network.
in reply to Jiří Eischmann
in reply to Jan Petrus
in reply to Jiří Eischmann
in reply to Jan Petrus
in reply to Jan Petrus
in reply to Jiří Eischmann
in reply to Jiří Eischmann
in reply to Koubik
in reply to Jiří Eischmann
in reply to Jiří Eischmann
Unknown parent
in reply to Jiří Eischmann
in reply to Miroslav Buček 🌳
in reply to Jiří Eischmann
in reply to Jiří Eischmann
in reply to Miroslav Buček 🌳
The media in this post is not displayed to visitors. To view it, please go to the original post.
Quite a milestone for Czech #Meshcore: Bohemia and Moravia functioned as two islands, because crossing the Bohemian-Moravian Highlands is not exactly easy. But today, the connection has been established. The two islands are beginning to function as a single network.
reshared this
Peter Vágner and Tissman reshared this.
I live exactly in the middle of the blind spot south west of Pardubice. Maybe it's time to join the network...
@jan_petrus someone has to start it in the area. 😉 A well-located repeater (on a hill, high building) can cover a large area and connect it with other regions.
We are quite high, especially north and east horizon is visible in long distance (can see Giant mountains from the bed actually). We will see 😉
@jan_petrus then you should also see the repeater in Orlické hory and a repeater with a solid antenna should reach it. The Meshcore mobile app has a useful feature: you can pick a spot on map, set how high the repeater would be and it will calculate signal coverage.
@jan_petrus they installed a new repeater south-west of Pardubice today and it has improved the coverage of the area quite a lot.
super thrilled to see the Boston Mesh developed live map system in the wild on Mastodon!
This entry was edited (Thursday, January 22, 2026, 5:59 PM)
I will jave ty try to connect from Poland side of Praděd.
yup that’s an idea … lets see, will have to check the receiver 😉
@zoufalec @jan_petrus if there is no repeater in your area, it won't work. Companions (the clients) have a reach up to 3km (5 in ideal conditions). It isn't enough to reach a remote repeater, especially in hills. You would need a repeater on a hill/high building/chimney with a good antenna and clear view to connect your area to the network. In the Meshcore app there is a useful feature: you can pick a spot on map, set how high the repeater would be and it would calculate the reach.
Why didn't you choose Meshtastic? It's much more widespread and fundamentally better for off-grid mesh communication.
@miroslavbucek because everyone around me is switching from MT to MC. I have no experience with MT, but from what I read and heard from others MT doesn't scale. I'm not really interested in keeping contact with friends in the same area, I could get a simple walkie talkie for that. What interests me is building a larger, country-wide mesh network. Hardly possible with MT, quite doable with MC. I can already chat with someone in Pilsen over MC. Good luck achieving that with MT. In practice MT can do 7 hops max, I've already received messages with 14 hops on MC, theoretically it can do 64.
@miroslavbucek IMHO MC community will be bigger than MT in Czechia this year if it already isn't. And it basically started here just last year. The shift is strong and quick.
I have a different use case: we often go into forests with the kids where there’s no cell signal. We have walkie-talkies, but their range is limited. For these purposes, Meshtastic is a better choice. The mesh can be formed by any node.
@miroslavbucek Yes, it's exactly the use case where MT is better than MC: ad hoc communication in one location. MC can do direct communication between clients, too, so in most cases it would work just fine, but e.g. in the scenario where you have three clients: A, B, C. A can reach B and B can reach C, but A cannot reach C. MT would work for communication between A and C and MC wouldn't.
2026-01-23 07:03:49
2026-01-23 07:03:49
2026-01-23 07:03:49
9588099
Peter Vágner reshared this.
FOSDEM 2026 (Brussels): Sun Feb 1, 09:35 CET, UB5.230 (my last talk)
Accessible Sovereignty: Why the Four Freedoms Depend on Inclusion (Community devroom)
Thesis: the Four Freedoms are not real if people with disabilities cannot use the tools, study them, share them, and improve them in practice. Accessibility is the operational layer that makes software freedom exercisable.
Peter Vágner reshared this.
2026-01-22 09:12:39
2026-01-22 09:12:39
2026-01-22 09:12:39
9585401
Peter Vágner reshared this.
The media in this post is not displayed to visitors. To view it, please go to the original post.
Rebuilding the indieweb one tendril at a time.
This is what we need to do. Rebuild the internet, reclaim the word social in social media and put people before platforms.
We must uncouple our attention from the tech-bro, elitist, narrative-controlling monopolies1 and get back to building individual, independant website presences. The indieweb.
Sites that are shared and aggregated by RSS feeds, collected into like-minded groups by webrings, boosted with decentralised social media platforms such as Mastodon, and cross-pollinated with interlinks and cross site conversations from individual posts.
Subvert, bypass, and starve the big platforms of their attention oxygen. Let them suffocate in a circle-jerk of AI generated slop.
There is now a growing movement of people dumping the scroll-gatekeepers, and building their own cheap, simple (often retro looking) independant websites.
Here is an example: Daryl Sun has a simple site that is packed with information and interesting links for anyone spending a little time to click rather than scroll (just like here).
In the spirit of re-wilding the web and creating interesting tendrils, I have updated my own ABOUT PAGE to let you know far more about me than you ever knew you wanted to know. It is a work in progress and will grow over time.
Another great example is Brennan, who writes some really interesting stuff and has recently begun moving his focus from posting on Medium to growing his independent site.
He always has tons of interesting outgoing links, is a member of a heap of webrings, as well as including a slash page of all his interests and projects.
I am gaining a lot of inspiration from this movement and urge you to consider dropping out of the shittosphere of big social media and spending that newfound scroll time creating your own independent online presence.
Let us tear the whole thing down and rebuild it fit for purpose.
If you already have a indie website drop me a link in the comments….I would love to follow and share.
- Facebook. Twitter X. Bluesky. Youtube. Tiktok. and the rest… ↩︎
#blackAndWhite #indieweb #photography #socialmedia
About
Everything you need to know about me: Who?A little about me.Contact.How to get hold of me.NOW.What I am up to in real time.Zen & me.My practice.Lens.My gear and a few of my favourite pics.AI &a…shojiwax.com
reshared this
Peter Vágner reshared this.
2026-01-23 20:46:47
2026-01-21 09:16:49
2026-01-21 09:16:49
9580764
Peter Vágner reshared this.
in reply to Mike Gorse
in reply to Peter Vágner
in reply to Mike Gorse
in reply to Peter Vágner
Today was nice; didn't have any new things demanding my attention, so just pushed a few existing things forward. Also managed to get back to looking at bugs.webkit.org/show_bug.cgi?i…. Found that, with my test, some text doesn't show up in the AT-SPI hierarchy if I just remove the conditionals mentioned in the bug. I'd really need access to the WebKit slack, since there is apparently some discussion there, but, when I try to join (from the Getting Started link on webkit.org), I get a message saying that the invite has expired. I wonder if the link works for other people. I emailed someone at Apple the other day; hoping that he gets back to me.
Peter Vágner reshared this.
@Mike Gorse Regarding webkit GTK a11y, is it know multiline content is sometimes trimmed to first line?
For example:
is only showing first to screen readers.
For example:
<p>
first
<br>
second
</p>is only showing first to screen readers.
@pvagner What I was describing is similar to that, but I was only seeing it with a patched local version. I was hoping that it might fix this evolution crash that I'm seeing when I try to authenticate a gmail account--that was really the thing that I'm most interested in trying to fix right now--but it seems not to make any difference. Anyway, is what you're describing something that you're seeing with an unpatched WebKit? If so, is there a bug open for it?
@Mike Gorse I can reproduce this all the time with epiphany and any tauri based app. I'm on arch linux with webkitgtk-6.0 2.50.4.
As for the bug reports I have no idea currently. There is loads of issues and I am not familiar with them.
As for the bug reports I have no idea currently. There is loads of issues and I am not familiar with them.
@pvagner I don't usually use epiphany, but I just tried it and found that it is very crashy for me--all I have to do is poke around www.google.com with Orca running. I don't know the code very well, but I'd like to try to investigate, since it seems like no one is working on WebKit's AT-SPI support right now, but now I'm being yanked away by unrelated day job stuff, so everything that's on my to-do list will have to stay there for now.
2026-01-21 09:15:40
2026-01-21 09:15:40
2026-01-21 09:15:40
9580774
Peter Vágner reshared this.
in reply to Devin Prater :blind:
in reply to Devin Prater :blind:
in reply to Khronos
Okay y'all, first public, alpha, version of Elmud! Basically, the install script should handle everything for you, if you're on Termux for Android, Linux, or Mac. Hopefully. If not, just load-file elmud.el and it'll load emacspeak-elmud.el too. If you need a screen reader, set up and use Emacspeak. For soundpacks, we have one automatically load for Erion MUD and Cosmic Rage. Those are the two I've found easily downloadable. So the client just downloads the sounds for you when you choose those two MUD's. You can also make your own gag or sound triggers. I'll put it up on Github soon but for now:
reshared this
Jonathan and Peter Vágner reshared this.
Are you saying that something like Emacspeak can be ran on a phone?
How do I go about installing this after I have Termux installed?
@khronos You'll need the Emacs package, so do pkg i emacs, then git clone github.com/tvraman/emacspeak, and then make inside the emacspeak directory, then get a speech engine working, I recommend just going with Espeak when starting out, make Espeak. If it doesn't work have AI fix it up. Then git clone ElMud, run the install script, and in Emacs, M-x ElMud.
GitHub - tvraman/emacspeak: Emacspeak --- The Complete Audio Desktop
Emacspeak --- The Complete Audio Desktop. Contribute to tvraman/emacspeak development by creating an account on GitHub.GitHub
Khronos reshared this.
2026-01-21 13:37:11
2026-01-21 09:12:20
2026-01-21 09:12:20
9580966
Peter Vágner reshared this.
My wife and I are fortunate to be in Iceland at the moment, and we managed to capture a timelapse of last night’s aurora. It was one of those moments in my life when I could hardly believe my own eyes. Enjoy!
#aurora #iceland #timelapse
The media in this post is not displayed to visitors. To view it, please go to the original post.
My wife and I are fortunate to be in Iceland at the moment, and we managed to capture a timelapse of last night’s aurora. It was one of those moments in my life when I could hardly believe my own eyes. Enjoy!
#aurora #iceland #timelapse
reshared this
Peter Vágner and Jason Fayre reshared this.
2026-01-21 09:02:07
2026-01-21 09:02:07
2026-01-21 09:02:07
9581414
Peter Vágner reshared this.
Canceled another subscription recently, the money that's left will probably go into donating to NTFY or other projects, another awesome software I seriously enjoy using.
Support your FOSS developers, even if it's just a small amount. Especially if you'd pay for the tool, like I did with pushover before.
Support your FOSS developers, even if it's just a small amount. Especially if you'd pay for the tool, like I did with pushover before.
Peter Vágner reshared this.
2026-01-19 21:48:00
2026-01-19 21:48:00
2026-01-19 21:48:00
9574241
Peter Vágner reshared this.
🗣️ Welcome Mia and Mil, two new voices designed to read texts in Luxembourgish using screen readers. They are making progress, but sometimes still struggle a little with Dicks' language.
🔧 Would you like to help them? Come and test them out, and let us know what you think: gd.lu/15pM9q
The media in this post is not displayed to visitors. To view it, please go to the original post.
🗣️ Welcome Mia and Mil, two new voices designed to read texts in Luxembourgish using screen readers. They are making progress, but sometimes still struggle a little with Dicks' language.
🔧 Would you like to help them? Come and test them out, and let us know what you think: gd.lu/15pM9q
Be part of the Luxembourgish screen reader adventure
This is the latest development at the start of the year: Mia and Mil, two new voices to reproduce texts in Luxembourgish as faithfully as possible. Come and see how you can help us get to grips with them.Digital accessibility portal of the Grand Duchy of Luxembourg
reshared this
Peter Vágner reshared this.
2026-01-19 21:45:14
2026-01-19 21:45:14
2026-01-19 21:45:14
9574692
Peter Vágner reshared this.
The media in this post is not displayed to visitors. To view it, please go to the original post.
Native Instruments did a 14-minute video on the accessibility features of their MK3 keyboards, barely enough time to even cover what the thing does, let alone the keyboard functions, not to mention it wasn't done by a blind user, the target audience.
I wasn't a fan, so I made my own video and I *still* didn't get through everything to do with it.
Anyway, here's a more than three times longer video at 48 minutes instead.
#InspiredBySound - Native Instruments Kontrol MK3 Accessibility Demonstration youtu.be/cMkm51Utijs
Audio download: onj.me/media/%23InspiredBySoun…
#InspiredBySound - Native Instruments Kontrol MK3 Accessibility Demonstration
In this video we take a tour of the Native Instruments Komplete MK3 keyboard.It's packed full of features and functions that were never accessible to blind u...YouTube
reshared this
Onj 🎶, Jakob Rosin, Borris and Peter Vágner reshared this.
2026-01-19 08:06:40
2026-01-19 08:06:40
2026-01-19 08:06:40
9566851
Peter Vágner reshared this.
haha. I have made blindsoft image describer-js (BSID-JS) open source. #opensource #cloudflare #workers #js #javascript github.com/averlice/bsid-js if you want to host or contribute at all, feel free. licensed under the gpl-v3 public distribution license.
GitHub - averlice/bsid-js: blindsoft image describer running on cloud flare workers
blindsoft image describer running on cloud flare workers - averlice/bsid-jsGitHub
reshared this
Peter Vágner reshared this.
2026-01-19 07:18:05
2026-01-19 07:18:05
2026-01-19 07:18:05
9571908
Peter Vágner reshared this.
Just gonna leave this here.
in reply to Andre Louis
in reply to victor tsaran
The media in this post is not displayed to visitors. To view it, please go to the original post.
Just gonna leave this here.
reshared this
Onj 🎶, modulux and Peter Vágner reshared this.
Is it a Postfelt piano I hear in the mix? :) Great piece, BTW!
This entry was edited (Sunday, January 18, 2026, 11:54 PM)
2026-01-19 15:30:43
2026-01-19 07:13:08
2026-01-19 07:13:08
9573121
Peter Vágner reshared this.
I should really go to bed but first have this thing I spent the last few hours making in Ableton. It's a remix of "Be Like You" by Taylor Acorn and it's probably one of my favorite songs of the last few years. Here's the original: youtu.be/Cys2K0rx2T0
The media in this post is not displayed to visitors. To view it, please go to the original post.
I should really go to bed but first have this thing I spent the last few hours making in Ableton. It's a remix of "Be Like You" by Taylor Acorn and it's probably one of my favorite songs of the last few years. Here's the original: youtu.be/Cys2K0rx2T0
Taylor Acorn - Be Like You (Official Audio)
Taylor Acorn - Be Like You (Official Audio)Stream Survival in Motion: https://onerpm.link/survivalinmotionFollow Taylor:▶ Spotify: https://open.spotify.com/a...YouTube
reshared this
Peter Vágner and johann reshared this.
2026-01-19 12:50:53
2026-01-18 23:17:47
2026-01-18 23:17:47
9567058
Peter Vágner reshared this.
in reply to Daniel Gultsch
in reply to Daniel Gultsch
Unknown parent
Unknown parent
in reply to Daniel Gultsch
Unknown parent
in reply to Daniel Gultsch
RE: gultsch.social/@daniel/1135490…
#Signal isn’t just AWS. It also has a hard dependency on Google’s push notification system (FCM) if you don’t want your battery to catch fire.
Signal’s attitude towards #UnifiedPush and #FDroid speaks volumes.
Peter Vágner reshared this.
Don't forget their policy on not allowing third party clients on their network.
wo wir gerade beim Thema #UnifiedPush sind.
Wenn conversations offline ist - z.b. beim Flugmodus - verschwinden die push Nachrichten. Sie kommen nicht verspätet sondern gar nicht.
Könnte man die nicht zum Zielserver ins dortige MAM schicken? oder wenigstens eine "verpasste Benachrichtigung" Mitteilung?
If I am not wrong, Signal does offer on their website, variants of their app without google's blobs implemented. I saw them on their github as well.
das klingt auf jeden Fall gut. Danke das du hilfst dieses tolle Projekt aus der Nische rauszuholen. 👍
Seit foss warn auch per unifiedpush warnt ist so ein resend durchaus wichtig geworden.
This entry was edited (Sunday, January 18, 2026, 9:24 AM)
What would you suggest as an alternative (not a facetious question - I'm genuinely curious)?
Great! All the best with the project. I've dropped you a follow so I can keep up with progress. 👊🏼
but it's a big step for most of the people!
However the problem is that most people don't even think about switching from WA to Signal - and NEVER heard about conversations #XMPP
#Schools could / MUST do more! Instead they are pushing students into digital dependencies. School leaders avoid the hassle to think, they appreciate conformity.
2026-01-19 08:38:35
2026-01-18 23:10:40
2026-01-18 23:10:40
9568879
Peter Vágner reshared this.
in reply to Darrell Hilliker 👨🦯♾️📡
reshared this
Peter Vágner and Tissman reshared this.
I’m extremely happy this is being developed. I worked with QLog and it’s self voicing interface for about half an hour yesterday and realized very quickly that using it on a regular basis would drive me completely insane!
2026-01-19 12:39:50
2026-01-18 22:59:36
2026-01-18 22:59:36
9569730
Peter Vágner reshared this.
I work as an audiobook quality controller. My employer uses ClickUp to manage tasks. Unfortunately, the web interface is unintuitive and inaccessible. It contains unnamed elements, menus that expand in all kinds of ways, and similar issues. I wrote to their developers, but even after years nothing has been fixed. Fortunately, ClickUp has an open API. So I used vibecoding and now I have my own minimalist HTML application that displays my tasks, start and end dates, comments, and attachments, and allows me to post comments. I still can’t change task statuses yet—we’ll see if I manage to solve that with the help of GPT. Of course, this is not how a blind person should function in an ideal world, but it is still a way we can help ourselves. That said, I still need a server where my PHP scripts run; it could probably be done with Python as well, but PHP seemed simpler to me since it’s already running on my server. BTW it would be ideal if Clickup api documentation is one simple document which I can throw to gpt but it seems that chatgpt already know how to use it.
Peter Vágner reshared this.
2026-01-20 20:21:36
2026-01-16 17:29:07
2026-01-16 17:29:07
9558030
Peter Vágner reshared this.
in reply to Paweł Masarczyk
in reply to clv1 has moved
in reply to Paweł Masarczyk
in reply to Paweł Masarczyk
in reply to miki
in reply to Paweł Masarczyk
in reply to miki
in reply to Paweł Masarczyk
in reply to Paweł Masarczyk
An accidentally discovered workaround that probably shouldn't even be mentioned as an actual one but got me through an hour-long webinar recording today: when an ad comes up on Youtube and your video is a part of a playlist, navigating to the next video skips it and loads the next one. As a result, if you skip to the next video when it's time to skip the ad and then return back to your video, the playback resumes. This obviously works also with remote controls on headphones.
reshared this
Peter Vágner and Zach Bennoui reshared this.
Thanks for the hint. In any case, I still suggest UBlock Origin.
I don't know; I have it on Firefox for desktop.
Not on iOS it seems? At least it didn't seem to work for me when I got trapped in those 2-hour-long "get rich quick" videos.
Huh, that's interesting. It always used to skip back to the beginning for me.
I have recently figured out that I can somehow still get a student YT premium subscription through my employee account, because SheerID apparently doesn't check permissions for us properly, so that's my workaround now.
@miki Just tried it on a video taken from the front page and it works here too as you can skip to the next suggested video from the player.
@MariahL I now use an ad blocker, but in Safari, if you encounter an ad, especially at the beginning of a video, refreshing the page seems to skip the ad. Almost every time I've tried it, this has worked for me without fail. I have no idea how nobody at google has fixed this. Unless somehow they can't.
good to know.Πs. if ayou have any vpn, try to connect with Birma and it will solve your problem too.
2026-01-15 19:04:21
2026-01-15 07:18:22
2026-01-15 07:18:22
9551043
Peter Vágner reshared this.
in reply to Jamie Teh
in reply to Jamie Teh
I recently had a need for a recursive C++ lambda that captures variables. This isn't possible in the intuitive way; you get:
error: variable 'someVar' declared with deduced type 'auto' cannot appear in its own initializer.
It turns out you can achieve this (albeit a bit horribly) by passing the lambda to itself! See this article for details: artificial-mind.net/blog/2020/…
error: variable 'someVar' declared with deduced type 'auto' cannot appear in its own initializer.
It turns out you can achieve this (albeit a bit horribly) by passing the lambda to itself! See this article for details: artificial-mind.net/blog/2020/…
Peter Vágner reshared this.
If you can use C++23, deducing-this makes it easier: dev0notes.com/intermediate/rec…
Recursive lambdas from C++14 to C++23 - Notes from /dev/null
There are algorithms that have natural recursive representation and are very cumbersome or painful to write iteratively. With the advent of lambda expression...www.dev0notes.com
Lol, I think this post encapsolates my distaste for C++ in a single go. Still fond of it, since it was my first programming language, but also, just ick :). Good luck, sir.
2026-01-15 07:17:01
2026-01-15 07:17:01
2026-01-15 07:17:01
9551087
Peter Vágner reshared this.
Victory adds Braille to its amps at no extra charge for blind players like @bermudianbrit thefretboard.co.uk/discussion/…
Victory adds Braille to its amps at no extra charge for blind players like myself - Amps Discussions on theFretBoard
For those that haven't seen this initiative: https://mixdownmag.com.au/news/victory-amplification-reveals-world-first-braille-guitar-amp/theFretBoard
reshared this
Peter Vágner reshared this.
2026-01-12 12:13:50
2026-01-12 12:13:50
2026-01-12 12:13:50
9536840
Peter Vágner reshared this.
AI Brailab Hungarian text-to-speech system Voice generator. That's right. I'm not done.
It's up on Jammable. Here's a cover of Brailab doing Dial Drunk, with Eloquence Reed. Enjoy.
jammable.com/brailab-hungarian…
an you tell Brailab from Eloquence in the song? I sure can. Poor Brailab. All squished because it only goes up to 10000HZ of audio quality. Aww awe, I almost feel sorry for it.
in reply to Tamas G
in reply to Tamas G
in reply to Mira
in reply to Tamas G
The media in this post is not displayed to visitors. To view it, please go to the original post.
AI Brailab Hungarian text-to-speech system Voice generator. That's right. I'm not done.
It's up on Jammable. Here's a cover of Brailab doing Dial Drunk, with Eloquence Reed. Enjoy.
jammable.com/brailab-hungarian…
an you tell Brailab from Eloquence in the song? I sure can. Poor Brailab. All squished because it only goes up to 10000HZ of audio quality. Aww awe, I almost feel sorry for it.
Brailab Hungarian text-to-speech system AI Voice Generator | Jammable AI Covers
Brailab Hungarian text-to-speech system AI voice & song generator. Create Brailab Hungarian text-to-speech system AI voice covers with advanced AI voice technology and 50,000+ AI voices.Jammable AI
Peter Vágner reshared this.
Any idea if you'll do a jammable bestspeech?
This is too catchy and funny. Hahahahaha. I want to try it now. How do you do it?
@tardis so on that page there's a "create" button, you press it and a dialog opens. There you click on the select song button, then you can upload from your computer. I think free accounts don't really exist anymore, or they might but they only let you preview. Otherwise it's $1.99 for the first month (unlimited cover credits) then $8.99 after.
How did you train the voice or how do you do that...
2026-01-11 07:28:55
2026-01-11 07:23:48
2026-01-11 07:23:48
9482296
Peter Vágner reshared this.
I had some free time today so I created this... audiopub.site/listen/13c92d44-…
Peter Vágner reshared this.
2026-01-10 13:56:44
2026-01-10 13:56:44
2026-01-10 13:56:44
9530930
Peter Vágner reshared this.
in reply to Viktor Strate
The media in this post is not displayed to visitors. To view it, please go to the original post.
I've just open-sourced a native macOS @matrix client that I've been working on! github.com/viktorstrate/mactri…
It is built with #swiftui and #MatrixRustSDK
GitHub - viktorstrate/mactrix: Native Matrix client for macOS
Native Matrix client for macOS . Contribute to viktorstrate/mactrix development by creating an account on GitHub.GitHub
reshared this
Peter Vágner reshared this.
2026-01-10 11:27:15
2026-01-10 11:27:15
2026-01-10 11:27:15
9530962
Peter Vágner reshared this.
The #blog post "A year of work on the ALPM project" has been released:
devblog.archlinux.page/2026/a-…
#ArchLinux #OpenPGP #RustLang #SovereignTechFund #STF #VOA #devblog
A year of work on the ALPM project
An overview of the work done on the ALPM project in 2024 and 2025.Arch Linux Dev Blog
This entry was edited (Sunday, January 11, 2026, 9:28 AM)
Peter Vágner reshared this.
2026-01-10 17:21:36
2026-01-10 11:01:39
2026-01-10 11:01:39
9530969
Peter Vágner reshared this.
Friends, this is the coolest thing I have ever seen and you should see it too. A guitar amp labelled entirely in braille. #accessibility #blind #blindness #braille
The media in this post is not displayed to visitors. To view it, please go to the original post.
Friends, this is the coolest thing I have ever seen and you should see it too. A guitar amp labelled entirely in braille. #accessibility #blind #blindness #braille
reshared this
Peter Vágner, Federico Mena Quintero and Seirdy reshared this.
2026-01-09 07:26:42
2026-01-09 07:26:42
2026-01-09 07:26:42
9522888
Peter Vágner reshared this.
The production on this track is insanely good. Love all the interesting ear candy and little details. On top of that it's just a really well written song. Jess Humphries - Spite
Peter Vágner reshared this.
2026-01-08 14:17:12
2026-01-08 07:34:29
2026-01-08 07:34:29
9514472
Peter Vágner reshared this.
Hey so Claude Code got DecTalk working in Termux on Android and got Emacspeak working with it. Very little lag. Emacs. Android. Org-mode. Bluetooth keyboard. Nov.el. Markdown-mode. Org-export. Calendar. Emacs. Mind blown.
#emacs #accessibility #android #blind #termux
reshared this
Peter Vágner and Oskar reshared this.
2026-01-14 17:05:06
2026-01-07 07:29:55
2026-01-07 07:29:55
9512538
Peter Vágner reshared this.
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Jage
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Jage
in reply to Bryn🥰
in reply to aryan
in reply to aryan
in reply to aryan
in reply to Bryn🥰
in reply to aryan
in reply to Bryn🥰
in reply to Bryn🥰
in reply to aryan
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Dennis long
in reply to Bryn🥰
in reply to Tech Singer
in reply to Bryn🥰
in reply to Tech Singer
in reply to Bryn🥰
in reply to Tech Singer
in reply to Bryn🥰
Unknown parent
in reply to Bryn🥰
Android users, are you ready?
Peter Vágner reshared this.
The more important question here is, is my terminal ready?
Now introducing, NVDARC for Android! First early testing version is available here: masonasons.me/softs/NVDARC.apk
reshared this
Jonathan, Pitermach, Tamas G, Andre Louis, Jakob Rosin, Marek Macko and Paweł Masarczyk reshared this.
oo what's this? More droid software for me to look at now, or when I get my Pixel! Should I wait to use this on the Pixel or should I use it now?
Part of my wants to wait because i'll be testing on a newer device
I wonder how this differs from the old Narwal? I'm not even sure if that one still works. Will need to try this.
@Jage this one has an on screen keyboard interface if you don't have access to a hardware keyboard, for starters. It also has the ability for you to remap hardware modifiers to other modifiers, I.E. remap right alt to insert. It also lets you change your TTS settings, which I'm not sure if Narwall did or not it's been forever. You can save machine profiles. Talkback accessibility actions are fully supported. More that I'm forgetting
@Jage it's basically the same as my iOS NVDARC app, minus a few things that Talkback/Android doesn't support or that I haven't figured out how to do yet.
@Jage Where can I find the iOS one? It doesn’t seem to be listed in the US AppStore, do you need a device with alternative app marketplaces?
Beautiful, thanks and will check it out.
@Jage The hardware keyboard support works the same as Narwhal, with the accessibility service. I figure why fix what ain't broke.
what's this coded in?
@Aryan I'm annoyed. Because while this does work, it's never going to be as smoothe as iOS. Talkback just doesn't have the same level of integration that VO does for certain things.
@Aryan accessibility scroll gestures and Magic Tap. The iOS app has it so that you can perform page scrolling gestures and send different keys that you can configure. I spent about 2 hours trying different things to make this work with Talkback and came up with nothing.
i will look into these and let you know if i find anything, but it would've been better if i myself had used features like this, it would have let me get better context.
@Aryan Either way though, like, feel free to test out the app as it is, it's still pretty usable I think
for sure i am gonna test this out
ok, what does magic tap do in the ios app?
Wow dude that is amazing. I have to try this.
This entry was edited (Wednesday, January 7, 2026, 9:06 AM)
wow, this is so awesome, might fire up the S22 just to install this on it and check it out for sure. Super great work on creating this, this is going to liberate me so much from my desktop when traveling.
There is no information on this. where can you find information on the project?
@Dennislong82 I don't have a page or anything available for it yet, as it's in early stages. Suffice it to say that NVDARC is a control app for NVDA Remote on Android and iOS.
I'm not going to start because, if I do, I will just go on and on and it'll get boring. I'm just going to say two things as shortly as I can. 1. I've been bugging people to update Narwhal, including one guy who has been good enough to put up with me here. The fact that I no longer have to do this and annoy people because all the work is done is beyond wonderful. 2. It works. I mean of course it does, you made it, but it does work. I'm sure I'll find issues and report them later. Note that I am using it on a very old phone, the Smartvision 3 running Android 11. It does work on that phone.
I've put this on two devices, as I said, a smartvision 3 running Android 11 and a Pixel running Lineage (Android 15). As for the SV, the speech and touch stuff seems to be unbreakable. Say all works with eloquence having run through a few book chapters. There are no pauses. Basically, I just can't make speech stop or command stop sending properly from the screen/the arrows. What doesn't work is the accessibility service. First, it didn't tell me about the service, I was hoping it would check whether it was enabled and, if not, say "your NVDARC service isn't available, sending hardware keyboard commands will not work, tap here to go enable it. Once you've enabled it, hit f11 in the application to forward keystrokes". I didn't need that myself only because you mentioned the accessibility service in passing. Anyhow, when I went and enabled it in accessibility settings, it said it was on. However, hitting f11 on the USB keyboard got nothing. I am still able to navigate the device, and still unable to send keystrokes directly to the remote machine. I can still tap/enter on any of the keys on the screen, and get those keys sent, but sending directly from the keyboard doesn't work. I may say that I have turned off Narwhal, Kapsys' own thing, and every other accessibility service except talkback. Is there a way to log what is going on or anything I can do to troubleshoot?
@techsinger How interesting. I might have to do some more testing here on my end for that, because others are saying they're having issues. Does this not work on one or both of your devices?
The Lineage machine is having a few more issues though I'm pleased to say that the service is working, at least it's forwarding keystrokes. First, though the service is working, TB is still capturing alt+everything. Alt+enter, for example, gets sent to the computer to grab properties, which is great, but it also taps whatever I'm on in the interface. Both TB and NVDARC seem to be receiving the keystrokes. Obviously, with most keystrokes, this doesn't matter, but when the key is a TB command, it does. Again, say all stops after a single sentence with Vocalizer and Dectalk. I am not able to use Eloquence on this thing at the moment, but hope to at some point. With both Vocalizer and DT, the response when typing is very slow indeed. I'm comparing it to the iOS responsiveness, which is quite good, in my respectful view. It's as if I'm sending sound over Sonobus/Teamtalk/Soundwire over the same network. Responsiveness on the SV3 is significantly better. If I may ask, what are you using yourself as a device? Also, do you think, from your experience designing it, that lower end but modern devices are going to help in terms of responsiveness? Again, thanks for all your work. These are teething troubles, but the thing has to have teeth to get teething troubles, and this thing definitely has teeth.
@techsinger Unfortunately, my only testing device right now is an S23+, which I'd say is a bit overqualified.
Sorry to bother you with more odd thoughts, but do you think controling an android phone possible, do you think, sending back stuff from TB?
no. Oh wait I'm not an android user.
2026-01-06 18:38:16
2026-01-05 20:23:00
2026-01-05 20:23:00
9509171
I have been notified about this #accessible text editor written in rust with a lot of interesting features.
GitHub - Ambro86/Novapad: Modern Notepad in Rust: read PDF/EPUB/DOCX, play MP3 audiobooks, create audiobooks with high-quality TTS, and manage bookmarks.
Modern Notepad in Rust: read PDF/EPUB/DOCX, play MP3 audiobooks, create audiobooks with high-quality TTS, and manage bookmarks. - Ambro86/NovapadGitHub
ondrosik reshared this.
2026-01-06 18:38:14
2026-01-05 19:58:32
2026-01-05 19:58:32
9505068
Peter Vágner reshared this.
in reply to menelion
in reply to Bryn🥰
in reply to Bryn🥰
in reply to Alex Hall
in reply to Bryn🥰
in reply to Alex Hall
in reply to Bryn🥰
A friend of mine, Beqa Gozalishvili, a very talented developer from Georgia the country, announced an early stage of his #SAPI5 wrapper for the popular #ESpeakNG #TTS engine. bug reports and feature requests are welcome, he says in his Telegram channel. He does speak English. github.com/gozaltech/espeak-ng… #Accessibility #ScreenReader #Windows #JAWS #NVDA
GitHub - gozaltech/espeak-ng-sapi: Sapi5 interface for espeak-ng text-to-speech synthesizer
Sapi5 interface for espeak-ng text-to-speech synthesizer - gozaltech/espeak-ng-sapiGitHub
Peter Vágner likes this.
reshared this
Paweł Masarczyk, Andre Louis, Pitermach, Peter Vágner and Bryn🥰 reshared this.
This is very cool! Glad eSpeak is back as a modern SAPI voice!
I believe @alexhall may find this interesting. IIRC you do use eSpeak as your primary, or am I mistaken?
@Bri I do. ESpeak is the voice I always come back to with NVDA. It would be really cool to have the NG variant as a SAPI voice.
@alexhall you can already download it. Pull the conversation if you can and check out the github page. I can get you a download link if you can't find it
@Bri I found it. I'll have to play with this sometime soon. My next few days are kind of booked, but I'm definitely intrigued.
@Bri I look forward to it. I'm using the Blastbay voices right now, but they still have that weird lag problem every so often.
2026-01-27 13:28:18
2026-01-05 19:46:27
2026-01-05 19:46:27
9506709
Peter Vágner reshared this.
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Landon
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to 🏳️⚧️PepperTheVixen🇵🇸
in reply to 🏳️⚧️PepperTheVixen🇵🇸
in reply to Devin Prater :blind:
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Devin Prater :blind:
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Amir
in reply to Sean Randall
in reply to Andre Louis
in reply to James Scholes
in reply to Sean Randall
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to James Scholes
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to James Scholes
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to clv1 has moved
in reply to Andre Louis
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Andre Louis
in reply to Andre Louis
in reply to Andre Louis
in reply to Luis Carlos
in reply to Andre Louis
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Chris Smart
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Zach Bennoui
in reply to Zach Bennoui
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Zach Bennoui
in reply to Zach Bennoui
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Chi Kim
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Jayson Smith
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Paul L
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Matthew
in reply to Matthew
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Matthew
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to shortwavesurfer2009
The State of Modern AI Text To Speech Systems for Screen Reader Users: The past year has seen an explosion in new text to speech engines based on neural networks, large language models, and machine learning. But has any of this advancement offered anything to those using screen readers? stuff.interfree.ca/2026/01/05/ai-tts-for-screenreaders.html#ai#tts#llm#accessibility#a11y#screenreaders
reshared this
Andre Louis, Sean Randall, Tamas G, Peter Vágner and Zach Bennoui reshared this.
Oh god, what if Chatgpt gets intergrateded into nvda?
I've started using eSpeak-ng. It's grating, but I can crank the speed up way higher than any other TTS I've ever used, especially the fancy AI shit that simulates breath draws and lip movement
@PepperTheVixen The reason it's grating is because unlike Eloquence and dectalk, Espeak only uses formant synthesis for the vowel sounds. For consonants and plosives, it instead uses concatenative recordings based on human speech. That's why even when you switch to a voice that sounds less sharp, the "t", "b", "p", and other sounds are still too sharp. This seems to be the primary cause of the fatigue most people experience while using ESpeak.
Lol just imagining an AI voice with lip smacking noises.
@pixelate@PepperTheVixen If you give chatterbox-tts an ASMR recording to clone, you can absolutely get it to make lip smacking noises.
Oh my goodness. Or even better, an AI voice chewing gum.
@pixelate@PepperTheVixen If you have a sample of someone talking while chewing gum, you can absolutely make that happen.
What an interesting read! Needless to say, I read it with Eloquence - LOL!
🇨🇦Samuel Proulx🇨🇦 likes this.
It's crazy that everyone is layering it in wrappers nowadays.
Do you know if codefactory are doing the same with their new android build?
Do you know if codefactory are doing the same with their new android build?
I sincerely hope someone will do the same for Orpheus. I'd even pay for it.
There is a 32-bit compatibility layer in the works for NVDA itself (although it currently only references SAPI4). But with any luck the need for every add-on to implement its own will go away.
github.com/nvaccess/nvda/pull/…
Support for SAPI4 via a 32 bit shim runtime by michaelDCurran · Pull Request #19412 · nvaccess/nvda
Link to issue number: Summary of the issue: 64 bit NVDA can no longer support SAPI4. Description of user facing changes: SAPI4 is again availble on 64 bit NVDA. Description of developer facing cha...GitHub
It does seem incredible to cut every 32 bit thing out so suddenly.
@cachondo@jscholes@FreakyFwoof@amir They don't have much choice. A lot of the libraries NVDA depends on are stopping 32-bit support this year.
I guess if this had happened a decade ago it'd have excited me. I'm obviously getting too old!
@jscholes@cachondo@FreakyFwoof@amir My understanding is that when this comes to addons, it's going to require some kind of secure addons API/layer. And it won't be ready for 2026.1, or maybe not even 2026.2.
Where are you getting the first part of that understanding from? I.e. the dependence on the secure add-on runtime. @cachondo @FreakyFwoof @amir
I see the "Secure add-on runtime" on the roadmap, with the note that "The first version of the runtime will provide support for speech synthesis and braille devices."
I don't see any implication that any 32-bit compatibility layer will only work for secure add-ons, which is hopefully a bit of a leap.
Still, the fact that people don't know what will or won't be happening, or whether their preferred synthesiser(s) will work or not, continues to be a big part of the problem. @cachondo @FreakyFwoof @amir
@jscholes@cachondo@FreakyFwoof@amir That's my assumption because the only things that really need a 32-bit compatibility layer are speech synthesizers and braille devices.
Regarding ESpeak-ng, AFAIC, the main complaint from users is its base tone, which cannot be solved by simply making new variants. In this regard, how about improving its MBrola voices?
🇨🇦Samuel Proulx🇨🇦 likes this.
@clv1@jscholes@cachondo@FreakyFwoof@amir The issue is that both of these are effectively concatenative or parametric, rather than formant, systems. So they will never be as intelligible as eloquence.
@FreakyFwoof@cachondo@amir You should be able to get either Gemini or Codex to help you, depending on what AI you have access to. The workflow would be:
1. download gemini-cli or codex-cli, and get them installed and configured.
2. clone all of the sourcecode from github.com/fastfinge/eloquence_64/3. Delete the tts.txt and tts.pdf files, so you don't confuse it with incorrect documentation.
4. Find any API documentation for orphius that's available, and add it into the folder.
4. Run codex-cli or gemini-cli, and tell it something like: "Using the information about how to develop NVDA addons you can find in agents.md, and the information about the Orphius API I've provided in the file Orphius-documentation-filename.txt, I would like you to modify the code in this folder to work with Orpheus instead of eloquence."
It will go away for five or ten minutes, ask you for permission to read and write the files it's interested in, and then give you something that mostly works. Now, build the addon, run it, and tell it about the errors and problems you have and ask it to fix them. In the case of errors, include the error right from the NVDA log, and for bugs and problems, tell it exactly what it's doing wrong, and exactly what you want it to do instead. Keep doing this until you wind up with a working addon.
Think of AI as a particularly stupid programmer, and you're the manager in charge of the project. You should be able to get this done without paying anyone.
GitHub - fastfinge/eloquence_64: Eloquence synthesizer NVDA add-on compatible with 64-bit versions of NVDA
Eloquence synthesizer NVDA add-on compatible with 64-bit versions of NVDA - fastfinge/eloquence_64GitHub
johann reshared this.
Well there's already a 32-bit addon for Orpheus floating about. I'd still rather pay someone competent to do it, even if they use AI. Proper programming terms would help narrow down the broken bits. I'm just an audio guy.
@FreakyFwoof@cachondo@amir Yeah, you can get AI to modify the 32-bit addon for you. That's how I got the first two eloquence prototypes; it helped me understand the problem and what approaches would work and what wouldn't. If you give it the 32-bit orphius addon, and the 64-bit eloquence addon, it should be able to understand the working approach to make an addon 64-bit, and make the modifications itself. The reason to give it the 64-bit eloquence addon as an example is so it doesn't decide to go down the GRPC route and include protobuf and a bunch of other nonsense.
Oh happy days 😊 that was the voice that used to come with the Hal ScreenReader isn't it? That was my first ScreenReader after my accident back in 1996 and I seem to remember the plug-in synth was called something like Apollo two or thereabouts such happy memories 🙂 but not really I used to sit up till about 4 am banging my head against the brick wall trying to figure it out but hey ho
And also for Kokoro and other speech synths.
@luiscarlosgonzalez@cachondo@FreakyFwoof@amir I didn't try Kokoro, because it cannot achieve a real time factor of 1 on CPU. By that I mean, to be fit for consideration with a screen reader, a text to speech voice must be able to generate one second of speech in one second or faster. In general, Kokoro takes two seconds to generate one second of speech. So it's not suitable.
What even about Tortoise or Cookie or that synth I don't know of
Supertonic works way better on my 2012 machine than Kitn. That one just cuts off things all the time and just doesn't sound good at all. I'm pretty sure it's not supposed to do that. Very weird.
🇨🇦Samuel Proulx🇨🇦 likes this.
Malware Bytes advises me not to visit this page, and that's after I paste the URL in my browser because Tweesecake doesn't recognize it as a URL. :)
Really interesting article. I'm particularly passionate about this subject, I've been fascinated with TTS for a number of years. I've trained many voices, both for Piper and some of the newer LLM based systems, and while I can't speak to the speed issue, training data is extremely important.
What you feed into these models has a big impact on the voice's performance overall. If you give it stuff scrape from the web, random audiobooks that weren't optimized for TTS, things like that, you're not going to get good results for the type of work screen reader users do every day. This applies to all of these systems, not even just neural networks. The latency / responsiveness issue is something we'll have to solve at some point, because I don't think using TTS systems last updated in 2003 is going to work out in the longterm, as much as I love Eloquence.
In my ideal world, we would have either a machine learning based or formant system that is easy to train / maintain. Big companies have lost interest in on device TTS, not even just for screen reader users. Many of the solutions being put out now are cloud based, and while developers are still creating on device models, as said in the article, they're not optimized for our needs and may never be. I think we have to take matters into our own hands and figure this out, but I believe with enough people we can make it happen.
🇨🇦Samuel Proulx🇨🇦 likes this.
@ZBennoui Agreed. I think blast bay is close to the right track. If only it was open and the issues pronouncing words were fixed. The speed and sound of the voices are top notch.
@ZBennoui We need a good formant system. Machine learning is useful for setting the model parameters. But I think the word to phoneme rules can’t be a neural network, because they have to be reproducible and modifiable. Even here though, machine learning could help though. I’d love a system where a user could submit a recording of a word, and the system could create the phonetic representation.
Yeah I completely agree, I happen to know Philip and have been talking with him extensively about his experiments with TTS. I can't go into a ton of detail, but I'll say what he said publicly. The system he's using is a hybrid approach of neural networks and formant synthesis, where he trains a model to output formant frequencies based on the audio data he feeds into it. I won't pretend to understand all the details, this is way above my pay grade, but as far as I understand this has never been done before by another developer.
This entry was edited (Tuesday, January 6, 2026, 6:13 PM)
@ZBennoui Yup. I just wish he wasn’t also trying to train his own phonemizer, because I really believe that has to be reproducible and modifiable for users. I’ve swapped multiple emails with him about an NVDA addon. But he’s pretty set on sapi for now until things stabilize both on the NVDA side and on his side.
@amir I too think we can make it happen by taking matters into our hands. I don't know how to code, but I'm at disposal to work on Portuguese language support, e.g. improving pronounciation rules, when time comes.
🇨🇦Samuel Proulx🇨🇦 likes this.
@FreakyFwoof The major problem with incorporating AI TTS into screen readers is latency. Maybe you can use it for say all, but it is not suitable for navigation. There are a couple of tiny TTS with low latency like Piper TTS, but the quality is not the best. Also multi lingual support and pronunciation for many uncommon words are issues for AI tts.
I assume you didn't mention the modern efforts with regard to DECtalk due to the legal situation with that source code being about as clear as mud?
@jaybird110127 Yup. That and the fact that the sourcecode isn't getting updated; just getting it to keep compiling is a huge effort. There is a 64-bit build, but it doesn't actually work. I consider dectalk pretty much dead, even though the source is available.
isn't it possible to "pregenerate" the speech with all the necessary IDs so that you can navigate and interrupt at will?
Just as one generates SSML from rich text (including maths formulas) before generating speech.
It would even be better to catch intonations, breaths and others, unchanged instead of letting the TTS generating a "pleasant full phrase" (a wrong expectation).
I find your post intriguingly close to the emerging reaction against the Ai-generated #mundaneslop 
.
@polx Maybe, but probably not. Doing that would result in a lot of wasted resources generating text I'm never going to listen to. Think about the average user interface: dozens of menus, and toolbars, and ads, and comments, and so on. Plus, the text changes constantly, on even simple websites. That's not even taking into account websites that just scroll constantly. It might be possible to create some kind of algorithm to predict the most likely text I'll want next, but now we've just added another AI on top of the first AI.
I think a better solution might be to make the text to speech system run on different hardware from the computer itself. This is, in fact, how text to speech was done in the past, before computers had multi-channel soundcards. This has a few advantages. First, even if the computer itself is busy, the speech never crashes or falls behind. Second, if the computer crashes, it could be possible to actually read out the last error encountered. Third, specialized devices could be perhaps more power and CPU efficient.
The reason text to speech systems became software, instead of hardware, is largely because of cost. It's much cheaper to just download and install a program than it is to purchase another device. Also, it means you don't have to carry around another dongle and plug it into the computer.
I have been developing a neural TTS system, focused on screen reading for many months, which offers instant responsiveness, but maintains good synthesis quality at the same time. And, BTW, it is not recommended at all to use espeak as a phonemizer backend as breaks the text embeddings during model training, especially if we use linguistic information. And, please consider to avoid overeading NVDA's python environment in your add-ons.
@rmcpantoja Yes, the only way to avoid messing with the NVDA Python environment would be to do an IPC server. But at that point, you're really just rewriting SAPI and it seems pointless.
@rmcpantoja The issue with not using Espeak is that it makes it impossible to have user dictionaries. When we use a neural network, linguistic rules are no longer deterministic. So it might say a word correctly with one voice, at one time, but not with another voice, or at another time. This makes it impossible for us to correct mispronounced words in a reliable way.
My results shows that a dedicated IPC server performs faster, E.G the synthDriver is ready for use in just 5 seconds; response is good even in longer sentences, but this can be attributed to the 4.2m model I'm using. And when I ran the model through a streaming vocoder, response is surprisingly realtime, suitable for screen reader. As for voice rate, I'm using a modification of the good "audiostretchy" pip package. I can't give more details ATM, but I hope this helps in your research
github.com/RHVoice/RHVoice
GitHub - RHVoice/RHVoice: a free and open source speech synthesizer for Russian and other languages
a free and open source speech synthesizer for Russian and other languages - RHVoice/RHVoiceGitHub
@shortwavesurfer2009 The problem with RHVoice is that it's concatinative (with some hidden markov model tricks for phoneme selection). So it will never be as easy to understand at extremely high rates of speed as eloquence.
2026-01-05 19:45:01
2026-01-05 19:45:01
2026-01-05 19:45:01
9506754
Peter Vágner reshared this.
For World Braille Day, I compiled a list of free braille art and tactile image libraries that I used for a project in one of my graduate school classes. This is by no means a complete list of all available resources for tactile graphics, rather it is a list of sources that were explored for this project. veroniiiica.com/free-braille-a…
Free Braille Art And Tactile Image Libraries | Veroniiiica
Free online resources for braille art and tactile graphics for blind and low vision students, with an emphasis on STEM content. Written for World Braille DayVeronica Lewis (Veroniiiica)
reshared this
Peter Vágner reshared this.
2026-01-05 19:50:09
2026-01-05 06:40:44
2026-01-05 06:40:44
9501170
Peter Vágner reshared this.
in reply to Amir
in reply to 🍂Melissa🌠
in reply to Amir
in reply to JamminJerry
in reply to Amir
in reply to Amir
in reply to Brandon
in reply to Amir
in reply to Brandon
in reply to Amir
in reply to Brandon
in reply to Amir
in reply to Amir
A new day, and a new accessible Telegram client for Windows, TAccess.
TAccess is a custom Telegram client built with Python and wxPython, specifically designed to be fully accessible for screen reader users (NVDA, JAWS, Narrator) on Windows.
github.com/mlapps88/t-access-a…
TAccess is a custom Telegram client built with Python and wxPython, specifically designed to be fully accessible for screen reader users (NVDA, JAWS, Narrator) on Windows.
github.com/mlapps88/t-access-a…
Release T-Access 1.0 · mlapps88/t-access-accessible-telegram-client
TAccess v1.0.0 - First Public Release We are proud to present the first stable release of TAccess, an accessible Telegram client designed from the ground up for efficiency and ease of use with scre...GitHub
reshared this
Peter Vágner, johann, Andre Louis, Paweł Masarczyk, Khronos and menelion reshared this.
It won't let me download it, and this makes me sad. :(
@EmeraldRose You may want to try this direct link, or paste it in your download manager if you use one: github.com/mlapps88/t-access-a…
wow! I broke this thing the first time I launched it! go me! it said there was an update available, and when I told it update now, it had an error.
Error dialog Update failed: 'UpdateDialog' object has no attribute 'main_frame'
now I can't close the program to restart it. lol! I guess this thing just isn't jerry proof. smile.
@JamminJerry LOL - maybe. But my main concern is the flurry of Python/wxPython apps that have popped up over the past few weeks, mostly vibe-coded. Vibe-coding isn’t necessarily bad, but it often doesn’t follow clean, well-established development standards. By installing such apps, we tacitly accept that we may be subject to all kinds of expected—and not so expected—issues. Now I'm not saying this particular app is vibe-coded. But these days whenever I hear Python/wxPython, I may cringe a bit.
Also, I find it weird that its on Github, but only the releases the code isn't actually there.
like it kinda defeats the purpose of putting it up on GH IMO.
like it kinda defeats the purpose of putting it up on GH IMO.
I couldn't get happy with anything I made, so good on them!
@serrebi Well, if anything, I'd really like to see your RSS reader succeed. It's still a bit resource-hungry here to be honest. But the idea of a fully accessible, full-text RSS reader is always enticing.
Is it? You must have a huge RSS library. I have a few hundred feeds and I think it's fine, but I'm using a nice desktop.
@serrebi I have tens, not hundreds of feeds. And the machine is also powerful with 32GB of RAM and a late 2024 modern CPU.
When does it use high resources. I'm not seeing it. I have 18 cores though, so maybe that's why. I care about this though. This is one of the reasons my format was a blessing, because now I have a clean slower machine to test on.
This entry was edited (Monday, January 5, 2026, 9:08 AM)
@serrebi Let me try your latest build then for a more accurate picture and report. I'm waiting for you to add the PR there about the full-text module excluding the first paragraph on some websites.
I'll do it now. Todays been crazy I just missed it.
@serrebi The latest versions should be significantly more optimized in this regard. I also experienced performance issues with Blind RSS, even on powerful hardware, which is why I tried to improve it—I hope successfully. 🙂
2026-01-05 20:42:00
2026-01-05 00:53:58
2026-01-05 00:53:58
9496961
Peter Vágner reshared this.
Remembering the birthday of Louis Braille, a reminder that the 3D printed buildings related to his life, his childhood home, the school where he studied and the Pantheon where he is resting is freely available, just contact me and I'll send you the file.
#LouisBraille
#France
#Braille
#ReplicaBuildings
The media in this post is not displayed to visitors. To view it, please go to the original post.
Remembering the birthday of Louis Braille, a reminder that the 3D printed buildings related to his life, his childhood home, the school where he studied and the Pantheon where he is resting is freely available, just contact me and I'll send you the file.
#LouisBraille
#France
#Braille
#ReplicaBuildings
reshared this
modulux and Peter Vágner reshared this.
2026-01-05 00:49:31
2026-01-05 00:49:30
2026-01-05 00:49:30
9497322
Peter Vágner reshared this.
here is my tribute to celebrate Braille audiopub.site/listen/3c9fba77-…
Peter Vágner reshared this.
2026-01-05 03:30:06
2026-01-05 00:39:45
2026-01-05 00:39:45
9499056
Peter Vágner reshared this.
Introducing Whack A Braille! A new Blind-first and a11y-first audio-based game aimed at increasing typing skills and #Braille literacy. Multiple game and input modes, fun sounds, and you earn tickets with each round that you can use to get silly prizes when you cash out at the Prize Counter! Practice your Perkins typing with the home-row setting and your grade 2 symbols and word signs. I'm still iterating, but enjoy this initial release. Go whack some moles! marconius.com/fun/whackABraill…
Whack a Braille!
Whack a Braille is a braille-first arcade game where fast fingers, sharp ears, and solid braille skills win tickets and prizes.marconius.com
reshared this
Peter Vágner and johann reshared this.
2026-01-04 13:09:10
2026-01-04 13:09:10
2026-01-04 13:09:10
9490707
Peter Vágner reshared this.
in reply to Tamas G
Check it out. Braille Display Calibration tool is now on GitHub for you all to admire in its full 96.4% C++ glory. It will get updates here from now on, but the private binary will also stay current.
github.com/tgeczy/BrailleDispl…
github.com/tgeczy/BrailleDispl…
Release Beta1 · tgeczy/BrailleDisplay_calibration
This is an initial release. It supports all major features I would want in a calibration tool. Best works with NVDA; JAWS still may have issues on multiline displays.GitHub
Peter Vágner reshared this.
Is that for the Cadence? or can it be used for other braille displays?
2026-01-04 21:59:57
2026-01-04 12:45:42
2026-01-04 12:45:42
9493889
Peter Vágner reshared this.
in reply to Bryn🥰
in reply to James Scholes
in reply to Bryn🥰
in reply to aryan
in reply to Bryn🥰
in reply to Peter Vágner
in reply to Bryn🥰
So I've just pushed a really early version of being able to control a Mac via NVDARemote to the NVDARC app, via a new Be controlled option in the settings page. Note/warning! Due to limitations in VoiceOver, the speech output is going to be less than desirable in a lot of cases, but at the moment, it's the best one can do. Thanks to @miki for igniting the spark in me to actually make this a reality. testflight.apple.com/join/edg8…
reshared this
Jonathan, Peter Vágner and modulux reshared this.
What method of VO speech capture are you using?
@jscholes Applescript. There's a get last phrase command you can send VO. That's the only way I've found short of coding a whole damn speech synth.
Yo what? That sounds crazy. Needa try this at some point.
@pvagner @jonathan859 Yeah, I wanted to bring NVDA Remote to Linux as well. Maybe that'll be my next project.
Now I wanna know how you did that, that's so awesome! Maybe we can finally make the NVDA remote protocol non-NVDA-specific?
This entry was edited (Sunday, January 4, 2026, 6:57 PM)
2026-02-25 05:44:39
2026-01-04 12:41:43
2026-01-04 12:41:43
9494555
Peter Vágner reshared this.
in reply to Daniel Gultsch
in reply to Khleedril
in reply to Daniel Gultsch
in reply to Daniel Gultsch
in reply to ♾️
Unknown parent
Unknown parent
in reply to Daniel Gultsch
Happy Jabber Day 🎂 🥳
On January 4, 1999, Jabber was first announced to the public¹.
Twenty-seven years later, Jabber—or XMPP, as it became known after standardization through the #IETF—remains the only truly vendor-independent, federated instant messaging platform.
In almost three decades, XMPP has never stopped evolving and remains our best tool for digital independence.
¹: tech.slashdot.org/article.pl?s…
#DigitalIndependenceDay #Jabber #XMPP #DiDit #DigitalSovereignty #DiDay #JabberDay
Open Real Time Messaging System - Slashdot
Jeremie writes "Jabber is a new project I recently started to create a complete open-source platform for Instant Messaging with transparent communication to other IM systems(ICQ, AIM, etc).tech.slashdot.org
reshared this
Nicoco, Gajim, Peter Vágner and Štěpán Škorpil reshared this.
That post contains more words and information than the article it refers to!
@khleedril Way back in 1999 it probably looked more like web.archive.org/web/1999041805…
These days jabber.org is just one instance among many, not a software project.
Jabber: Home
Jabber is a new open-source instant messaging platform designed to be open, fast, and easy to use and extend.web.archive.org
So, what is IRC?
datatracker.ietf.org/doc/html/…
"Jabber—or XMPP, as it became known after standardization through the #IETF—remains the only truly vendor-independent, federated instant messaging platform."
RFC 1459: Internet Relay Chat Protocol
The IRC protocol is a text-based protocol, with the simplest client being any socket program capable of connecting to the server. This memo defines an Experimental Protocol for the Internet community.IETF Datatracker
We all just pretend the Matrix protocol with federation, builtin E2E encryption and several open source client and server implementations does not exist.
@hisold As far as I know, Matrix locks you into their own encryption. Multiple XMPP clients supports OTR, besides OMEMO.
Federation? So many people sticks with matrix.org.
Amount of stable/actively developed server can be comparable here, so Matrix doesn't seem to be superior compared to XMPP.
Amount of clients suddenly doesn't make a protocol good. Maybe make one if you disagree with all of the clients that currently exists?
Federation? So many people sticks with matrix.org.
Amount of stable/actively developed server can be comparable here, so Matrix doesn't seem to be superior compared to XMPP.
Amount of clients suddenly doesn't make a protocol good. Maybe make one if you disagree with all of the clients that currently exists?
This entry was edited (Monday, January 5, 2026, 8:55 AM)
Element only develops the Rust and JS SDKs. And the libOLM crypto library of course. All other SDKs are community projects.
On the server side, there is Conduit which forked into conduwuit which forked into Tuwunel which will eventually reach maturity, if the moon is in the right phase. The phone number mapping service is controlled by the foundation. You could use alternative ones or none.
I think Matrix is getting there eventually.
As far as I can tell, work on the standard has slowed down due to being more and more complete. This is good for projects implementing it.
I think the entire ecosystem is on a good path. It needs fair competition between companies and community projects. Let's wait another 2 years and see what we get.
It's time for the old protocol to cough up some gold coins
Pitermach
in reply to Nolan Darilek • • •