Peter Vágner reshared this.
mastodon - Link to source

The first release of My Photo Library Described is out, for Mac and Windows. Open your photo library, arrow through your photos, and hear each one described by the AI built into your computer: Apple Intelligence on the Mac, Windows AI on Copilot+ PCs. Nothing is uploaded. Free and open source.

theideaplace.net/introducing-m…

#Accessibility #ScreenReader #VoiceOver #JAWS #NVDA #AltText #AppleIntelligence #CopilotPlus

Introducing My Photo Library Described: Your Photos, Described as You Browse, on Mac and Windows

Since I started working on my Image Description Toolkit (IDT) back around January of 2025, I’ve had a goal: open my photo library on Windows or the Mac and have descriptions read automaticall…
The Idea Place
#Accessibility #screenreader #AltText #nvda #jaws #voiceover #appleintelligence #copilotplus

reshared this

Peter Vágner reshared this.
glitchsoc - Link to source

My Zoom F3 talks now.

As a blind field recordist, I've modded the F3's firmware so it speaks: every menu item as you scroll, which option is selected, popups like "SD Card Full!", confirmation dialogs, and even the date and time setup. No sighted help needed.

Would you want a step-by-step guide, with no technical knowledge needed, to put this on your own F3?

#Accessibility #Blind #ZoomF3 #FieldRecording

  • Yes, I own an F3 and want this (0%, 0 votes)
  • Yes, I’d consider getting an F3 for it (11%, 2 votes)
  • Just curious, following along (76%, 13 votes)
  • I’d want this for other recorders (11%, 2 votes)
17 voters. Poll end: Thursday, November 12, 2026, 9:27 PM

#Accessibility #blind #fieldrecording #zoomf3

reshared this

in reply to ozo 🐻🍉
glitchsoc - Link to source

Casey Reeves (it/its)

@sergiodeprado I know right? I've been super busy making the actual patcher and such so that people can literally try this for themselves. I’ve officially gotten the few remaining items handled, including the support for all 7 languages.
Of course, I can only evaluate english and french myself, so I cannot be 100% certain that other languages are handled correctly. Still, they are now hopefully usable, and not just random words strung together!
@ozo 🐻🍉
Peter Vágner reshared this.
mastodon - Link to source
If you speak German, which I certainly don't but thanks to AI now I do, this thread might be of use on the SSI263. forum.classic-computing.de/for…

Nachbau für SSI-263? - VzEkC e. V.

Hi, es gibt ja Projekte für alle möglichen Retro-Chips, die heute nicht mehr erhältlich sind... CPUs, SIDs, PLAs etc.
VzEkC e. V.

Peter Vágner reshared this.

Peter Vágner reshared this.
mastodon - Link to source
Well. I asked about private NVDA remote servers this morning. Since I wasn't quite happy with what I found, here's my take on it. Feel free to try, share, give feedback and all that ;)
github.com/thyhellcat/nrps
#NVDA #NVDARemote #A11y #Rust

GitHub - thyhellcat/nrps: Private relay server for NVDA Remote Access

Private relay server for NVDA Remote Access. Contribute to thyhellcat/nrps development by creating an account on GitHub.
GitHub
#a11y #rust #nvda #nvdaremote

Peter Vágner reshared this.

in reply to Peter Vágner
mastodon - Link to source

Chris Eilers

@pvagner You're most welcome. Yeah, I saw the Go one while making mine, but it has some security issues, memory leaks etc.
Please do let me know how you like that one. It should indeed be very simple to set up, but that's always harder to judge for the creator.
@Peter Vágner
Peter Vágner reshared this.
mastodon - Link to source
And well, since I'm back on the nerdy side of things:
Ever had that one bloody shortcut that just doesn't want to reach the intended program? Maybe it's being intercepted by something else first!
Try and find out what that "something else" exactly is.
github.com/thyhellcat/hkdetect

GitHub - thyhellcat/hkdetect: Find out which program owns a global hotkey on Windows

Find out which program owns a global hotkey on Windows - thyhellcat/hkdetect
GitHub

reshared this

Peter Vágner reshared this.
mastodon - Link to source
If you have ever wanted to set up a headless Debian server from scratch over SSH (without needing to plug in a keyboard or a monitor), I wrote a guide. Mostly for my future self if/when I have to do it again. ctemm.me/headless-debian-insta…

Installing Debian completely headless over SSH

How to build a custom Debian installer USB that starts an SSH server on boot, so you can install without a monitor or keyboard.
Carter Temm

reshared this

Peter Vágner reshared this.
mastodon - Link to source
Got LineageOS 23.2 booting on my iPhone 7. Currently trying to get audio to work, since right now I have to use a fake TTS service that forwards Backtalk's speech over USB to Orca. @TheQuinbox I blame you for putting me in this rabbit hole, without Backtalk I probably wouldn't have started this. Look what you've done!
@Quin

reshared this

Peter Vágner reshared this.
mastodon - Link to source

ssi263-speech 0.8 is out!
The biggest release yet of the emulated SSI-263 (Votrax SC-02) speech chip, running the real units' own firmware.

- One NVDA add-on for every unit: the Braille Lite 2000, the Slovak Braille 'n Speak 2000, the Speak-Out, the Accent SA, the Accent-mini and, new, the Apple II's Mockingboard. It carries over your 0.7 settings and offers to remove the old add-ons.
- The Braille Lite 2000 now speaks 19 languages: 17 new voices, from Arabic and Hungarian to Russian and Turkish, thanks to an anonymous donation of Blazie's language updates. Screen readers switch to them by language.
- The chip sounds closer to the real one: fuller vowels, grittier fricatives, and clean BASIC tones.
- The emulator: every language in its Firmware menu, programs and files loaded straight into flash, Blazie's own firmware update, and a native BT Speak and BT Braille package.
- Android imports every voice at once from an NVDA add-on, and the Apple app is in TestFlight for iPhone, iPad and Mac.

Thank you to Astra, the donors, and everyone who tested, reported and listened.

NVDA add-on:
github.com/tgeczy/ssi263-speec…

SAPI 5 voices:
github.com/tgeczy/ssi263-speec…

Blazie emulator for Windows:
github.com/tgeczy/ssi263-speec…

Android app:
github.com/tgeczy/ssi263-speec…
testflight.apple.com/join/cJYh…
Please find Linux, SSInger and more on the release page:
github.com/tgeczy/ssi263-speec…

Join the SSI-263 Speech beta

Available on iOS
testflight.apple.com

reshared this

Peter Vágner reshared this.
mastodon - Link to source

The BrailleDB database brailledb.seeinghands.org/ contains features and specifications of, to our knowledge, every Braille display, notetaker, Braille keyboard, and similar device available for purchase today. We also include many devices which have been announced but can’t be purchased yet and some historical devices that may still appear in the used device market. With BrailleDB, you can find a full list of features for any device, see a side-by-side comparison between any set of devices, and filter devices by the features you are looking for. Because there are a lot of things to consider about any access technology, we also have a full glossary of terms explaining what each feature is and why you might consider them important in making your decision.

Devices can be filtered by many attributes and as few or as many as you desire. Whether you want to see all multiline displays, all 40-cell displays that are available for purchase today with the ability to edit text files, or all displays shorter than 20 centimeters in length that have onboard translation for Spanish and Japanese, BrailleDB can do that.

What’s coming next?
This initial version of the database is only the beginning. We have a few ideas for things to do. The used Braille display market is strong, given the high price of new equipment, so there is a benefit in adding more historical display information here. We already have 12 discontinued devices listed, but there are far more that we should consider adding.

Mostly, though, what’s next is to see what we never thought of. As you use the database, if there’s anything you think would be useful, use the contact link on the top of any page to let us know. The same goes for anything that doesn’t work right or where something could be made easier to use. We know a database can never answer all the questions that come up when choosing between options, but if you have one that would fit, we are going to try to find a way that it can.

Get involved
Would you like to help improve this system with us? We would appreciate anyone interested in joining our team of volunteers. You don’t need any particular skills, just an interest in what we’re doing. If you’re interested in testing new features and letting us know where we’ve gotten it wrong, describing devices as we add them to the database, discussing how to implement new suggested features, or any idea you have for how you could help, we would be delighted to welcome you to the team. If this sounds interesting, fill out this quick volunteer survey, s.seeinghands.org/brailledbvol… and we’ll soon be in touch to discuss what you would like to do.
seeinghands.org/index.php/2026…
#Braille #Blind

New Braille Device Database – Seeing Hands

New Braille Device Database
Seeing Hands
#blind #braille

reshared this

Peter Vágner reshared this.
mastodon - Link to source

Has anyone noticed NVDA no longer announces virtual desktop names on recent Windows 11 builds?
After some simple debugging on Win11 25H2 (build 26200.9550), Explorer stopped exposing desktop names via the #32769 nameChange event NVDA monitors. It now fires a UIA SelectionItem_ElementSelected event under DesktopsList, which existing code doesn't handle.

#NVDA #ScreenReader #a11y #Accessibility #Windows11 #UIA

#a11y #Accessibility #Windows11 #screenreader #nvda #uia

Peter Vágner reshared this.

in reply to cary-rowen
mastodon - Link to source

NV Access

Glad you got it fixed! There ARE builds numbered 26200.9550 AND 26300.9550 - so we're both probably reading correctly - I'm on the regular stable branch, are you perhaps on a preview branch out of interest? Or maybe it's just different region branches or something.
Peter Vágner reshared this.
glitchsoc - Link to source
so, because I was disappointed with microsoft's key remapping implementation, partly because of how limited it seems to be and partly because I had to just come down from a future platform release to an earlyer insider build just to try this, have nopilot! this tool properly converts your copilot key to right control. NVDA sees it as right control, not windows+shift+f23 like microsoft's implementation, or many AHK scripts. the project is at github.com/aryanchoudharypro/n… and the direct download is at github.com/aryanchoudharypro/n…

GitHub - aryanchoudharypro/nopilot: Turns the Copilot key on Windows laptops back into a real Right Ctrl key, properly.

Turns the Copilot key on Windows laptops back into a real Right Ctrl key, properly. - aryanchoudharypro/nopilot
GitHub

reshared this

in reply to aryan
mastodon - Link to source

Munchkinbear

problem is, it's running here and sadly, at least on my HP omen gaming laptop it's still reporting windows+shift+f23. Yes, before you ask, it's running, I installed it per the docs and I've rebooted. Yes, before you ask this, I've ran it as an administrator as well. It *acts* like the control key but it's still showing as windows+shift+f23 with both NVDA and JAWS.
Peter Vágner reshared this.
glitchsoc - Link to source
The Microsoft Neural Voices Manager and Microsoft Neural Voices Driver add-ons for NVDA have been updated. If you use these add-ons to enjoy high-quality offline neural voices, this is definitely an update worth getting. And yes, for the first time, I also have two offline Neural Persian voices!
Here are some highlights from the changelog:
• Added support for new languages, including Amharic, Azerbaijani, Bengali, Bosnian, Persian, Filipino, Gujarati, Armenian, Indonesian, Javanese, Georgian, Kazakh, Khmer, Kannada, Lao, Macedonian, Malayalam, Mongolian, Marathi, Malay, Maltese, Burmese, Nepali, Pashto, Sinhala, Somali, Albanian, Sundanese, Swahili, Tamil, Uzbek, Zulu, and others.
• Added support for new countries and regions corresponding to the newly supported languages.
• Added 33 new languages, bringing the total to 77, and 25 new countries, bringing the total to 105.
• Added 94 new voices, bringing the total to 349 available voices.
Download links - pages are in Russian and Google Translate with Chrome makes it easier:
NeuralVoiceManager:
nvda-addons.ru/page.php?id=Neu…
NeuralVoices Driver:
nvda-addons.ru/page.php?id=Neu…

NeuralVoiceManager (Управления нейронными голосами Microsoft) NeuralVoiceManager дополнение для NVDA

Дополнение NeuralVoiceManager версии 2026.07.07 (Дополнение NeuralVoiceManager для NVDA предоставляет удобный интерфейс для управления нейронными голосами Microsoft.Вы можете просматривать голоса с фильтрацией по языку и стране.
nvda-addons.ru

reshared this

Peter Vágner reshared this.
mastodon - Link to source
Now announcing my new app, FastRoute. This one is a long time in the making. The idea of FastRoute is that you have sessions, and within these sessions you have different blocks that you can add, in any order you want, to send audio anywhere you like and process it however you like. This app is for Windows users, and is designed to create a similar ish experience to something like Audio Hijack from the Mac. Input blocks include things like input devices, applications, system audio (with exclusions support), internet streams. Effects include a 4 band parametric EQ, a 10 band EQ, channel mixer, pitch shifter, normalizer, compressor, and more. Outputs include output device, recorder, etc. FastRoute can also send and receive audio across the network, using it's net send/net receive blocks. We have our own protocol, but also support VBAN (VB Audio Network) protocol as well! This barely scratches the surface of the power of FastRoute, so check it out! github.com/masonasons/FastRout…

Release Latest Build · masonasons/FastRoute

Automated build from commit 5956773 Version: 0.6 Commit: 5956773 Built: 2026-10-09T18:30:14-05:00 Vendor the ASIO SDK and host one ASIO device Enumeration reads the registry, so it is safe to call ...
GitHub
in reply to Bryn 😻
mastodon - Link to source

Alex Hall

This is kind of awesome! I almost want to do an audio demo, just because of how easy it now is to capture my voice and NVDA. I was going to ask you if there are plans to let users record each source block individually, but then I realized that I can just add a record block after each source, with another record block at the end of the chain to get all the sources together.
Peter Vágner reshared this.
friendica (DFRN) - Link to source
just testing Spacetrace by @Derek Lane. I tryed it in musical context. github.com/dgl1984/spacetrace

@Derek Lane

reshared this

in reply to ondrosik
mastodon - Link to source

Derek Lane

I tried to make one control do that. Something always fell apart when I did. Either something would look bad visually, in VDA would read the wrong numbers, and the sliders in tone trace and space trace turn in to edit fields when you press enter on them, and adding context sensitivity on top of the questionable stack made a mess. c
Peter Vágner reshared this.
mastodon - Link to source

Hi! A quick survey for screen reader, speech synthesizer, and Braille display users in the open-source communities.

I'd like to know what assistive technologies you use. Fenrir? Speakup? Orca with a speech synthesizer and/or a Braille display? BRLTTY?

Do you use the virtual console, desktop environments such as Xfce, GNOME, KDE, or a combination of these?

The goal of this survey is to gather feedback and ideas from the community and help provide the best possible accessibility solutions for FreeBSD as well.

Thanks for sharing your experience!

#FreeBSD #Accessibility #ScreenReaders #BrailleDisplay #OpenSource #UNIX

#Accessibility #opensource #freebsd #unix #screenreaders #brailledisplay

reshared this

in reply to Alfonso Siciliano
mastodon - Link to source

Cleverson

Fenrir has evolved well; I like it when I want or need a pure console environment. Speakup is good as well, but maybe more difficult to port, since it's too tied to the Linux kernel. Its advantage is, it support more hardware synths, which is good for sysadmins, but those pieces of hardware are becoming rare nowadays. Orca is the only "de facto" option for desktops; it works well under mate and gnome, and kinda works under XFCE and KDE. the only issue I'd mention is, I feel that the entire stack isn't that stable. Sometimes the environment freezes, i.e. both speech and keyboard stop working, but I have no idea what component it has to do with, maybe the interaction of at-spy with GTK. As for synths, the most popular is by far Eloquence, but it's old, proprietary and 32-bit software. Some implementations try to emulate it in a open manner, e.g. OpenEVV. the most convenient synth to distribute in live images, for system install and admin, is ESpeak, which supports a lot of languages, but few people use it regularly, as it has too metalic a voice. Besides it, there are others like RHVoice (sounds more human), and TGSpeechBox, which is new and few people use until now, but I consider a promising option.
Peter Vágner reshared this.
glitchsoc - Link to source

Ever wanted to be able to control your Android screen reader with JavaScript? Backtalk now lets you do just that! @Aryan has already made scripts to combine the more actions and go to channel buttons in the YouTube app into accessibility actions instead, thus making videos only one swipe element, and doing the same for the update and view details buttons in the app updates section of the play store. Here's how simple this is, speak a message whenever settings opens:
export const manifest = {
id: 'hello',
name: 'Hello',
version: '1.0',
author: 'You',
minApiVersion: 1,
testedApiVersion: 1,
apps: ['com.android.settings'],
permissions: ['speech'],
};

backtalk.onAppEnter(() => backtalk.speak('Hello from a script'));

@aryan

reshared this

in reply to Quin
mastodon - Link to source

Elijah Massey

Do scripts support manipulating Braille? It would be very cool to have something like Braille Access in TalkBack, although I don't think it would be a good idea to just do what Apple does and have a limited set of applications in this environment that cannot be added to by plugins, and I'm not sure it makes sense to distribute an application environment like this with the screen reader itself. But if it could be done with scripting, and people could download a Braille Access add-on and plugins with applications people have written for it, that would be amazing I think. The plugins could even create entirely new, Braille first interfaces for Android applications.
in reply to Quin
gotosocial - Link to source

the esoteric programmer

interesting! do we have some kind of documentation on what's available? also, what about sandboxes? we wouldn't want scripts to be able to run forever and overload backtalk or something like that. Also, is there a way to distribute scripts to people?
Peter Vágner reshared this.
mastodon - Link to source

Here's a wacky Linux factoid--

If you look at the RAID UUID in the superblock on disk it's in little-endian byte order:

Linux Software RAID version 1.2 (1) UUID=cccdcc9b:f86a0b98:98df7949:8f2f7bb8 name=localhost:boot level=1 disks=2

But if you look at the metadata for active volumes in the /run/mdadm/map file, the UUIDs are in big-endian (network byte order) format:

md126 1.2 9bcccdcc:980b6af8:4979df98:b87b2f8f /dev/md/localhost:boot

I spent some time today being confused by why the UUIDs weren't matching before I realized what was going on.

#Linux #RAID #DFIR

#linux #raid #dfir

Peter Vágner reshared this.

in reply to Hal Pomeranz
mastodon - Link to source

Bruce Elrick 🇨🇦

Since endianness is a relationship between byte order and word order, doesn't little endian effectively encode word size sequence of a composite structure like a UUID?

It looks like the above just blindly treats it as 4 x 32 bit words, each stored little-endian, which doesn't match any of the UUID schemes in terms of internal structure.

Weird.

Peter Vágner reshared this.
mastodon - Link to source
The Tech-VI announcement list celebrates its sixth anniversary today. It began on October 10, 2020. My reasons for starting it actually began in the 1990s, when I was working for Blazie Engineering. I was keeping up with blindness assistive technology even then. I had been a technology enthusiast since the 1970s when I was exposed to products as low tech as talking clocks and 4-track tape recorders and as high-tech as the Optacon, the Speech Plus Calculator and the Kurzweil Reading Machine. In addition to selling their own products, Blazie Engineering became a one-stop shop for various blindness products, including DOS and Windows screen readers, Braille translation software, speech synthesizers and Braille embossers. I was keeping up with updates to some of those products and forwarding them to my co-workers at the company. How many of them cared enough to read all of those emails I’ll never know but they never asked me to stop so I kept emailing them updates.
In 1998, I began working at AbiliTech, where I was providing technology training to blind children and adults. Most of my co-workers were doing the same type of work and I continued to keep up with various types of blindness technology. I once again took it upon myself to send out regular updates to my teammates on products like JAWS, Connect Outloud, Kurzweil 1000 and other products.
In the late 1990s, Amy Ruell ran a wonderful list called Amy’s Filters and Forwards. Amy was subscribed to many mailing lists and she used that list to forward relevant announcements to her subscribers. Those messages were related to blindness and blindness technology. At one point, Amy began charging a small annual fee for the service, which I gladly paid. I even considered reaching out to her to ask if she’d like assistance in collecting announcements for the list in exchange for me being a part of the business. As she and I had never made contact at that time and she would have likely had no idea of who I was, I decided against it.
In 2002, I moved to Associated Services for the Blind, now VisionLink, where I ran their computer training program. During my 14 years working at that agency, I continued to send my co-workers updates on all types of software. Nobody asked me to stop doing it. I never received a visit from HR, warning me to stop spamming the fourth floor staff.
Even when I moved to Comcast in 2016, I continued to send some updates to members of my team related to Web accessibility and software that we used.
I was also sending various updates to other mailing lists. For the most part, they were received positively but I eventually wondered if it was time for this type of service to have its own dedicated list. I knew about and had great respect for Top Tech Tidbits and I had probably been reading their newsletters since they first started. However, what I wanted to do is to carry on the work started in Amy’s Filters and Forwards, where announcements would be sent individually as separate emails, rather than combining them into a weekly newsletter. I briefly thought about charging an annual fee for the service. I decided against it for several reasons and I stand by that decision. I didn’t know if enough people would be interested in subscribing to a list where they might receive a substantial amount of emails on certain days dealing with blindness technology but the only way to find out was to try it. This is how the Tech-VI list was born six years ago on October 10.
One thing that I had to decide on was the name for the list. I wanted a name that would indicate that this was a blindness technology list. My original name was VI-Tech. I was quite happy with that name and would have chosen it but at the time there was a discussion list called BlindTech. I didn’t want people to confuse the two and so I felt that I had to come up with another name. I reversed the name and came up with Tech-VI. In all honesty, I didn’t care for the name at all. It felt very unnatural but, for whatever reason, I went with Tech-VI for the list’s name.
As of today, we have 1955 subscribers and the subscriber count continues to grow. I am still amazed, pleased and humbled that so many of you trust me to provide you with this information and I take that trust very seriously.
At this point, I’d like to offer a few reminders.
1. Please remember that I do not write most of the emails that I distribute. They come from a variety of sources and I always try to ensure that the original source link is provided in the message. If I have my own comments to offer regarding a specific post, I always place it at the top of the message, prefaced with the line “From the Tech-VI list owner.”
2. If the message is something that I wrote, I will clearly put my name in the subject field, such as “From David Goldfield’s Blog.”
3. While this list is and always will be free of cost, I am setting up services for those who wish to send monetary donations. Donations will never be required or even requested but they will always be appreciated. You’ll be able to donate using either Ko-fi or PayPal. I’ll alert all of you once that has been set up.
4. While this list has a focus on blindness technology, I am not available to provide free email support or free training. If you’d like to discuss paid training opportunities, feel free to visit my training page on my Web site. You can also email me with inquiries at info@DavidGoldfield.com. I am also open to discussing possible employment opportunities where my skill set can be utilized; serious offers only, please. Feel free to email me at info@DavidGoldfield.com.
5. If you’d like to browse the Tech-VI archive going back to the very first message from October 10 of 2020, you can do so via this link.
6. Feel free to let others know about the Tech-VI list. I don’t have the time or resources to engage in promotion or advertising. The subscription address is tech-vi+subscribe@groups.io.
7. Finally, I would like to thank all of you who have chosen to subscribe to this list, whether you just joined this week or you’ve been with me for the past six years. Your loyalty and trust are so appreciated. Also, if you are inclined to do so, I will always appreciate your prayers for me and for my family.
I wish you God’s blessings and peace.

reshared this

Peter Vágner reshared this.
mastodon - Link to source

Tired of VoiceOver reading "box drawings light horizontal" over and over in Claude Code, Antigravity, Pi and other TUI apps?

I made vo-art-squash, a small open-source Mac command-line tool. It adds VoiceOver pronunciation entries that silence the 160 box-drawing and block characters (U+2500–259F) used for frames and borders, in all apps. Easy to undo,
and it backs up your prefs first.

github.com/vick08/vo_art_squas…

#VoiceOver #Accessibility #Blind #macOS #A11y
Tested on Mac 26.x only.

GitHub - vick08/vo_art_squash: Silence decorative box-drawing characters from TUI apps in macOS VoiceOver

Silence decorative box-drawing characters from TUI apps in macOS VoiceOver - vick08/vo_art_squash
GitHub
#a11y #Accessibility #blind #MACOS #voiceover

reshared this

in reply to patricus
mastodon - Link to source

Kaveinthran

@patricus For NVDA users, you can achieve something similar without installing an add-on. Just add a regular expression to NVDA’s speech dictionary.

Go to NVDA menu → Preferences → Speech Dictionaries → Default Dictionary → Add.

Use the following settings:

Pattern: [\u2500-\u259F]+

Replacement: Leave empty

Type: Regular expression

This should silence the same 160 Unicode box-drawing and block characters targeted by vo-art-squash.

If you also want to suppress some Powerline symbols commonly used as decorative terminal separators, you can use this expanded pattern instead:

[\u2500-\u259F\uE0B0-\uE0C8]+

One caveat: the Default Dictionary applies globally, so these characters will also be silenced outside terminal applications. You can experiment with NVDA’s Temporary Dictionary first.

This doesn’t address terminal redraws or other accessibility issues, but it offers a simple way to reduce unnecessary speech. I haven’t tested it extensively across synthesizers and applications, so I’d be interested to hear how well it works for others.
#nvda

#nvda @patricus
in reply to Kaveinthran
mastodon - Link to source

victor tsaran

@kaveinthran @patricus Yep, I mentioned just as much in the thread above. THis is why I am not yet sure whether an add-on is a good thing. It may be if I add a few things and configuration options, otherwise, I will look silly submitting an add-on with a one-liner regular expression.
I think I have ideas though.
@patricus @Kaveinthran
Peter Vágner reshared this.
glitchsoc - Link to source
Hello all,
my new audio game, convenience express: convenience store simulator, is entering a public beta testing phase. this is an AgNow exclusive title during the testing phase, because it's the most convenient way to get testing reports, and helps promote ag now, so a win win.
this is a convenience store simulation game, with lots of twists and chaos as well. run your very own corner store, keep the books balanced, keep the store stocked with its over 2500 products, scan people's ID for cigarrette and alcohol sales, survive inspections, prevent drunks from knocking down whole shelves, fight off thieves with a fly swatter or let police handle them, or just chill in the main menu and watch a demo store play out without player input. it's up to you.
download AgNow from agnow.brynify.me, create account, and play convenience express (agnow exclusive alpha) from the games library. please make liberal use of ctrl+shift+r in the game to report any issues, and most importantly have fun!
please boost for reach!

reshared this

friendica (DFRN) - Link to source
Helped a friend to upgrade two debian based servers with #ispconfig3 and #seafile from Debian 11, over Debian 12 to Debian 13. Moved from php 7.3 / php 7.4 to php 8.4 / php 8.5 and from python 3.9 to python 3.13 in the process.
#ispconfig3 #seafile
Peter Vágner reshared this.
mastodon - Link to source
Clipman server can now be accessed via the web (opt-in of course) so if that's your jam, update your servers accordingly. New users, check out onj.me/clipman.

Clipman - Accessible clipboard history

Fast, searchable clipboard history across desktop and mobile, designed for keyboard and screen-reader use.
onj.me

reshared this

Peter Vágner reshared this.
glitchsoc - Link to source
After many weeks and 4 games I am finally launching on Patreon! patreon.com/oriolgomez

Oriol Gomez — Mods and tools to add blind accessibility to videogames

Get more from Oriol Gomez on Patreon. Mods and tools to add blind accessibility to videogames. Support Oriol Gomez and get exclusive access to their work.
Patreon

Peter Vágner reshared this.

Peter Vágner reshared this.
glitchsoc - Link to source

Windows testers wanted: an early alpha of AAE, the Accessible Android Emulator, is ready to try on Windows.

AAE lets blind people create, run and test Android virtual devices without sighted help. Every device has a screen reader from its first boot, takes your keyboard, and plays its audio through your computer.

This is a very early build. It hasn't been run by anyone but me yet, things will break, and parts of the Mac app are still missing.

What's there:
- Setting up Google's emulator and tools for you, no Android Studio needed
- Creating, starting, stopping, copying and wiping devices, with Backtalk or your own screen reader
- Device mode: your whole keyboard goes to Android, Alt Tab and the Windows key included, while NVDA and JAWS keep their own keys. Control Windows Escape brings it back
- Gesture mode: screen reader gestures from the keyboard
- Installing apps, the clipboard both ways, screenshots, the device's volume
- AAE's messages spoken through NVDA, JAWS or Narrator
- Updates itself, so you'll be offered the next stable release

Not there yet: the accessibility inspector, speech log, device log, shell, apps and accessibility services windows, snapshots, battery and location. The aae command line, installed alongside the app, does all of these meanwhile. Open a command prompt in AAE's folder and run aae help.

You need Windows 10 or 11 on an Intel or AMD processor (not ARM), with Windows Hypervisor Platform turned on in Windows features. AAE isn't signed yet, so SmartScreen may warn you: choose More info, then Run anyway.

Download the installer:
github.com/aaron-gh/accessible…

Release notes and the Mac build:
github.com/aaron-gh/accessible…

If something goes wrong, Help, Save Diagnostic Report makes a text file you can read before sending. Tell me which screen reader you use, and whether AAE's messages were spoken. Replies here or GitHub issues are both fine.

#Android #Accessibility #A11y #ScreenReader #NVDA #JAWS #Blind #Windows

Release Development build (1c3d73f) · aaron-gh/accessible-android-emulator

Development build 5 of AAE, made from the latest change on master. It's for testing, before a stable release, and is replaced by each new change. Built from 1c3d73f: Reorganise the README: getting ...
GitHub
#a11y #Accessibility #android #blind #Windows #screenreader #nvda #jaws

reshared this

in reply to André Polykanine
glitchsoc - Link to source

aaron

@menelion Deleting a device doesn't delete the Android version it was made from. AAE keeps them separately.

A device is like your own phone: its settings, apps and data. An Android version (Google calls it a system image) is the copy of Android that devices are made from. Each one is around 2 GB and is shared by every device made from it, so it stays when you delete a device.

To delete an Android version, open the File menu and choose Android Versions (Control Alt A). Choose the version and press Delete. AAE won't delete a version while a device still uses it, so delete those devices first.

To check for yourself, paste these into File Explorer's address bar:

%LOCALAPPDATA%\aaron-gh\AAE\data\devices
Your devices, one folder each.

%LOCALAPPDATA%\aaron-gh\AAE\data\sdk
The emulator and tools AAE downloaded. The Android versions are in its system-images folder.

If you have Android Studio, AAE may have used its folder instead, and the Android versions are in %LOCALAPPDATA%\Android\Sdk\system-images.

Logs and downloads are in other folders inside %LOCALAPPDATA%\aaron-gh\AAE\data.

Uninstalling AAE leaves all of this in place. To remove everything, delete %LOCALAPPDATA%\aaron-gh\AAE. The emulator also keeps small settings files in %USERPROFILE%\.android, which you can delete if you don't use Android Studio.

@André Polykanine
Peter Vágner reshared this.
glitchsoc - Link to source

AAE 0.1.0 is out, the first release of the Accessible Android Emulator: a Mac app for creating, running and testing Android virtual devices without sighted help.

It sets up Google's Android emulator and tools itself, so you don't need Android Studio, and downloads the Android versions you choose, showing Google's licence for you to accept. You can have several devices of the same version under different names, each with its own apps and data.

What it does:
- Every device has a screen reader from its first boot. On Android 8 and later, AAE can download Backtalk for you, or you can install your own.
- Your whole keyboard goes to Android, with Command as Meta, including shortcuts like Command Space. VoiceOver keeps its own keys, and Control Command Escape always brings the keyboard back to the Mac.
- Gesture mode does screen reader gestures from the keyboard: swipes, two-part swipes, double taps, multi-finger gestures, and holds that last as long as you hold the key. Tab moves the touch point from item to item, so you can tap a particular control in an app with no screen reader running.
- Speech and sound you can rely on. AAE checks the device can speak on every start, and repairs it or installs eSpeak NG if it can't. It plays the device's audio itself, without the emulator's crackle, and raises the pitch of older Android versions, which play slow in the emulator.
- Testing tools: an accessibility inspector with basic checks, a speech log of everything the screen reader said, the device log filtered by app, and a shell.
- Install apps by copying them in Finder and pasting into AAE, choosing which devices to install on. AAE asks before turning on an app's accessibility service or keyboard, and remembers your answer. New screen reader builds can go on every device at once, keeping their settings.
- Snapshots with names and notes, the clipboard both ways, battery, location, text messages and calls.
- It updates itself.

It's for Macs with Apple silicon. AAE isn't notarised by Apple, so the first time you open it, go to System Settings, Privacy & Security, and choose Open Anyway.

github.com/aaron-gh/accessible…

#Android #Accessibility #A11y #ScreenReader #TalkBack #Blind #macOS #softwareDev #development

Release AAE 0.1.0 · aaron-gh/accessible-android-emulator

The first release of AAE, the Accessible Android Emulator: create, run and test Android virtual devices without sighted help. Sets up Google's Android emulator and tools itself, and downloads Andr...
GitHub
#a11y #Accessibility #android #blind #talkback #screenreader #development #MACOS #softwaredev
Peter Vágner reshared this.
mastodon - Link to source

Is this someone's solution to the terminal bug? Cool if it's a replacement for the removed feature. Interesting approach to addon naminng for sure.

NVDA Fighter: The Immortal RPG Guardian Shield & Necromancer
An epic, immortal guardian shield for NVDA. Protects NVDA from terminal output bombs (>10,000 chars) and speech engine crashes, enforces High Process Priority and Win32 crash-modal suppression, performs alchemical memory trims, and deploys an immortal out-of-process Shadow Beast daemon that automatically revives NVDA if it is ever slain.
Note have not looked at the repo yet.
github.com/ashu-choudhury/nvda…

GitHub - ashu-choudhury/nvda-fighter

Contribute to ashu-choudhury/nvda-fighter development by creating an account on GitHub.
GitHub

Peter Vágner reshared this.

Peter Vágner reshared this.
mastodon - Link to source

I've published a new, general-purpose terminal specification that gives any program a way to tell the terminal what it's doing: idle, working, waiting, finished, or failed, and why. It is easy to implement on both sides. I've written more about why: mitchellh.com/writing/program-…

The problem this solves is generic and applies in compelling ways to all sorts of use cases (e.g. package managers, build tools), but it really rears a hideous head when it comes to AI and the "agentic inbox" problem.

I found over 250 different agent orchestrators (with various goals) that all individually implement a heuristic based approach to detecting whether things like Claude Code are working, blocked, done, etc.

If you look at Herdr’s commit history, you can see around 10 compatibility changes to detect Claude Code within just the past 3 months. (This isn't a dig at Herdr! They're doing great work with what they have, and everyone else is doing the same thing).

Heuristics are not the way. And over-indexing on agentic use cases isn't the way, because this problem is really interesting for other tools too: Homebrew, Terraform, Cargo, etc. etc.

Proprietary protocols and out of band APIs are not the way. They have an O(N) integration problem and require extra work to work over SSH or within local VMs/containers.

The program status specification (OSC 7501) solves this in a reliable, deterministic way that is general-purpose and can be easily adopted by the entire industry.

Read more: mitchellh.com/writing/program-…

A Terminal Protocol for Program Status (OSC 7501)

Mitchell Hashimoto

Peter Vágner reshared this.

Peter Vágner reshared this.
pleroma - Link to source

This is gonna be awesome. Iroh is great, this is probably where we should be heading

iroh.computer/blog/iroh-global…

Iroh global content discovery

How to discover content globally with iroh using BitTorrent mainline DHT.
www.iroh.computer

Peter Vágner reshared this.

in reply to x0
mastodon - Link to source

Andre Louis

@x0 My SC-88 official demos never played properly, so I don't know if it was supposed to be able to handle it via Secret Source or not, but now mine very definitely does. Try it with the old one and see, then mine.
It also handles the Mu2000 plugins, both mine and the upstream build without VL. Lots of Yamaha midi files play weirdly with the normal FooMidi, work better with mine and sound more correct.
@x0
Peter Vágner reshared this.
glitchsoc - Link to source

Wel, guess you won't have Paperback for iOS for at least another week, because Apple doesn't allow developers to use nouns as their app names, and then takes a week to tell you that. Seriously if I hadn't sunk time into it and have people who really want it I'd throw in the towel on this fucking platform. I'm so tired of this.
Guideline 4.1(a) - Design - Copycats

Issue Description

The app's metadata contains third-party content similar to a popular app or game already available on the App Store, from a developer's website or distribution source, or from a third-party platform. This creates a misleading association with another developer's app or intellectual property.

Specifically, the app's description and name include references to Paperback.

This entry was edited (Tuesday, October 6, 2026, 2:56 PM)

reshared this

in reply to Quin
hometown - Link to source

Melissa Roe

That's nuts. I wish people would communicate these things more succinctly when working with these systems. It turns out there's a fingerprinting error with the Colorado Department of Education, which they haven't fully communicated to their license applicants unless you scream in their ear about it. Let alone having a messy website that sometimes doesn't load webpages properly and providing the wrong web address for submitting a contact request, sometimes I wonder if we shouldn't just let AI take over because the humans in power are gatekeeping, stupid assholes. I hope things continue to improve. You are a phenomenal developer, and I really love what you've done with Paperback so far. I can't wait for it to be on IOS. Best of luck with everything.
Peter Vágner reshared this.
mastodon - Link to source
The media in this post is not displayed to visitors. To view it, please go to the original post.

A recording of the talk “Jabber/XMPP: 25 Years of Digital Independence” that I gave at Linux Days in Prague the other day is now available online.

Learn why open source is necessary but not sufficient for digital sovereignty and why we need open standards like XMPP.

gultsch.video/w/mHRCPQ3SYTUio8…

#XMPP #Jabber #DigitalSovereignty #DigitalIndependence

Daniel Gultsch

2026-10-05 05:45:25

peertube - Link to source

Jabber⧸XMPP: 25 Years of Digital Independence


This talk was recorded at LinuxDays in Prague.

When it comes to digital independence, various solutions are being discussed in the field of communication and instant messaging. On the one hand, there is the centralized service of a US non-profit organization. On the other hand, there is the publicly documented API of a company that was largely built with venture capital. However, it might be worth taking a look at long-standing solutions that follow an actual internet standard and stand out due to numerous independent implementations.

Jabber/XMPP is one such internet standard and is "ready for prime time".

In over 25 years, the XMPP instant messaging protocol has seen its fair share of ups and downs. This talk highlights the recent history of the protocol and explains why open source alone, without federation and vendor-independent standards, is not enough for true digital sovereignty. The current status quo of XMPP is showcased through modern features like emoji reactions, moderation, read state synchronization across multiple devices, and time zone indicators for contacts. Additionally, the presentation offers a look ahead at upcoming protocol extensions and the initiative to return to the IETF with "XMPP 2.0" to update the existing RFC standards. Finally, attendees will learn how to easily and smoothly get started in the XMPP ecosystem today.

PDF Slides

Recorded by AVC SiliconHill, 2026


#xmpp #Jabber #digitalsovereignty #digitalindependence

reshared this

Peter Vágner reshared this.
mastodon - Link to source

Somewhat randomly, I published mp3dec today, a safe Rust no_std mp3 decoder, ported by hand from minimp3. I've actually been noodling at it for over a year as a relaxing side project. At some point, I hope to run it on RP2350.

In any case, I hope people find it useful or at least interesting.

reshared this

in reply to Matt Campbell
mastodon - Link to source

Raph Levien

@matt Thanks! Yes, there are a few things that are SIMD (arch-specific) in the original. But my little microcontroller doesn't have SIMD, so the motivation for me isn't as strong.

Actually that's one reason to actually commit the test suite :)

@Matt Campbell
Peter Vágner reshared this.
glitchsoc - Link to source

Shortcut Menu 0.5 is out, making safe mode safer.

Safe mode now always gives your screen reader back. If it ends without one turned on, by Close, by the screen turning off, or after 30 seconds with nothing touched, it turns your screen reader back on, and tells you which one. It also checks its own voice works before turning anything off.

Thanks to everyone who tried 0.4 and reported back.

github.com/aaron-gh/shortcut-m…

#Android #Accessibility #A11y #ScreenReader #TalkBack #blind

Release Shortcut Menu 0.5 · aaron-gh/shortcut-menu

Shortcut Menu 0.5 Safe mode never leaves you without a screen reader unless you choose that. If it ends without one, by Close, by choosing a service that is not a screen reader, by the screen turn...
GitHub
#a11y #Accessibility #android #blind #talkback #screenreader

reshared this

Peter Vágner reshared this.
glitchsoc - Link to source

Shortcut Menu 0.4 is out, with safe mode.

If your screen reader or your phone's speech stops working, open the menu with the accessibility shortcut and press volume up three times quickly. Every screen reader turns off, and the menu speaks for itself with its own built-in voice, SVOX Pico, so you can drag your finger to find a screen reader and turn it back on.

github.com/aaron-gh/shortcut-m…

#Android #Accessibility #A11y #ScreenReader #TalkBack #Blind

Release Shortcut Menu 0.4 · aaron-gh/shortcut-menu

Shortcut Menu 0.4 Safe mode: if your screen reader or your phone's speech stops working, press volume up three times quickly in the menu. Every screen reader turns off, and the menu speaks for its...
GitHub
#a11y #Accessibility #android #blind #talkback #screenreader

reshared this

in reply to aaron
mastodon - Link to source

Simon Jaeger

Yeah, it's not quite a direct comparison because the same problem never needed to be solved for Windows; we have an entire keyboard with assignable shortcuts. But the sentament was there, I think.
I also always forget that the Windows 7 utility Manager wasn't speaking through Narrator initially, since it uses the same voice.
Peter Vágner reshared this.
mastodon - Link to source
To my hungarian friends and followers! Also to @Tamasg. The rhvoice.hu site has been updated. I encourage you to test new hungarian RHVoice language module with three voices: imre, Anna and Katalin. Uninstall the old ones before installing these new versions.
@Tamas G

reshared this

in reply to Tamas G
mastodon - Link to source

Zvonimir Stanecic

Try it on long hungarian versus short hungarian vowels spelling, and on the hungarian text with exceptions. I tried to implement the most of espeak goodies. Human told me that it is not possible to parse espeak's exceptions directly to foma. Claude 5.5 did it.
Peter Vágner reshared this.
mastodon - Link to source
A guide and a small toolkit for writing programs for the Blazie Engineering Braille Lite 2000 notetaker.
Mising a few things, but was able to figure out most of the available functions.
github.com/jage9/blazie-dev-gu…

GitHub - jage9/blazie-dev-guide: Guide and toolkit for writing programs for the Blazie Braille Lite 2000 notetaker

Guide and toolkit for writing programs for the Blazie Braille Lite 2000 notetaker - jage9/blazie-dev-guide
GitHub

reshared this

Peter Vágner reshared this.
mastodon - Link to source
I was showing my 6 year old an old DOS car racing game from 1990 called Stunts which I used to play as a kid when I had some vision. He was having a little trouble with the race track editor and I couldn't quite remember all the UI. So... I used this opportunity to see how useful AI could be as a real-time accessibility assistant for a game. I installed chatGPT work on that machine and then connected to that chatGPT work session from chatGPT Work on a separate machine beside him. (Note I did it remotely so that the game would always stay in focus). Then, I was able to ask chatGPT to describe the screen at various points, and even get it to perform a few actions / mouse clicks. Its understanding of the UI from screenshots was quite amazing really, and even most of the mouse clicks achieved what we needed. But, the big annoyence today really is the length of time it takes from asking a question / asking it to perform an action, to actually achieving the end result. As there is a lot of thinking by the cloud model, then using the computer skill to screen capture, then more cloud thinking, then computer skill to move the mouse, then probably more screen captures, and more cloud thinking etc. So a query / action could take 30 seconds to a minute. And in some cases, that may be too long for the game and it may go back to its intro screen, especially if you're on the main menu for example. To be clear I don't expect this to be true realtime in terms of playing the game, and I also fully understand why today it does take that amount of time. However, I did feel it was an interesting experiment all the same. And at the end of the day, it did allow me and 6 year old to independently play the game. He is actually getting really good at it now and understands the UI, so no more need for AI there.

reshared this

Peter Vágner reshared this.
glitchsoc - Link to source

It's finally here: Paperback 1.0! My fast, fully accessible ebook and document reader is now out on Windows, macOS, Linux, and Android, with iOS following the second Apple decides to actually look at it. The changelog is once again way too big for any post, so here are the highlights:
New platform: Linux! As an AppImage or a tarball, with desktop integration so documents open straight from your file manager.
New formats:

  • Comic book archives (.cbz).
  • M4B and MP3 audiobooks, split into their chapters.
  • Man pages, both man and BSD mdoc, gzipped or not.
  • reStructuredText.
  • Windows Write (.wri) files. Thanks, Mastodon.
  • WinHelp (.hlp) files. Yes, really.
  • Word 6 and Word 95 documents.

Big new stuff:

  • OCR for scanned PDFs, using the engine built into Windows and macOS. Press Enter on a scanned page, or batch OCR a whole range of pages with Ctrl+Shift+O.
  • MathML in EPUB and HTML is now rendered as AsciiMath via MathCAT. Jump between formulas with M and Shift+M, and open the original MathML in a formula view with enter or space.
  • Find All, listing every matching line so you can jump right to the one you want.
  • Tables, Lists, and Pages views in the elements dialog.
  • Mark a spot with Alt+F9 and copy everything from there to where you are with Alt+F10, no more shift-arrowing through 40 pages.
  • Audio book playback speed, from half speed to 3x.
  • Any document can now be opened in the web view.
  • Headings are drawn at a size matching their level.
  • Relative Go To: +n and -n in Go to Line, Page, and Percent.
  • Heading navigation from the table of contents for EPUB, MOBI, and CHM books with no headings of their own.
  • A Ukrainian translation.

Plus a mountain of fixes, especially to PDF (I still hate PDF), MOBI/AZW3 books, which are now much faster and no longer cut off at 20 MB, FictionBook, and large books in general.

And of course, the mobile apps. They open every format the desktop does, with seamless read aloud using your choice of voice, background audio book playback, navigation by headings, pages, links, tables and more, a sleep timer, export, and keyboard shortcuts that match the desktop. They're $19.99, once, per platform, and buying them is what pays for the work on Paperback. The desktop versions stay free.
Desktop downloads: paperback.dev/downloads
Android: play.google.com/store/apps/det…
Full changelog: paperback.dev/readme.html
Huge thanks to everyone who tested, filed bugs, translated, and put up with me during this development cycle. Enjoy! #Paperback #accessibility #blind

Paperback - Apps on Google Play

Accessible and blazing fast ebook and document reader.
play.google.com
#Accessibility #blind #paperback
in reply to Quin
mastodon - Link to source

Jonathan Mosen

Am I just missing it in the iOS beta, or is there no integrated Bookshare support? I find being able to browse, search for, and download Bookshare books first in Voice Dream Reader, and now Dolphin Easy Reader, very helpful. An additional Easy Reader advantage is content syncing across Windows and iOS. So if I add a document at work, I can pick it up on my phone when I have a chance. Does Paperback accommodate syncing including reading position?
in reply to Jonathan Mosen
glitchsoc - Link to source

Quin

@JonathanMosen You can blame no bookshare on Benetech not being willing to play ball. Been yelling at them to get me an API key for V2 since 0.6.0. It currently doesn't sync reading position, but I plan to add a service to do this for a small monthly fee, or that you can host yourself, eventually.
@Jonathan Mosen
in reply to Jonathan Mosen
glitchsoc - Link to source

Quin

@JonathanMosen No, I've just gotten silence. Reached out to their contact email and got nothing, then had a benetech engineer open a few paperback issues and asked him. He gave me a work email, which I've emailed thrice over the span of a few months, and gotten nothing.
@Jonathan Mosen
in reply to Quin
glitchsoc - Link to source

Quin

@JonathanMosen I can't help but feel like benetech just doesn't want people using their new API, which is quite the departure from 10+ years ago when QRead got bookshare support. But the SensePlayer supports the V2 API, so maybe I just need to be an AT company.
@Jonathan Mosen
in reply to Quin
mastodon - Link to source

Robert Kingett

Book publishers, yes. Technically it is an SCP , (a secure FTP) feed that the publisher can upload titles to. I am indie published, so I asked for access to the server but haven't heard back at all, yet. I haven't even asked about audiobook uploads yet. @TheQuinbox @JonathanMosen
@Quin @Jonathan Mosen
in reply to Jonathan Mosen
glitchsoc - Link to source

Johann

@JonathanMosen Something about them still working on V2 of their API?
I actually tried reaching out to my local Bookshare contact here and they forwarded me to Bookshare South Africa of all places and told me exactly that they're closing new API access for now until they have v2 ready.
@Jonathan Mosen
in reply to Jonathan Mosen
mastodon - Link to source

Elijah Massey

@JonathanMosen Regarding syncing, there is no syncing of reading position yet as Quin said, but I think syncing the actual books is better, since on both iOS and Windows, Paperback just opens books from your own folders and doesn't store them in hidden folders that are hard to access and sync like VoiceDream Reader does. So you could install Syncthing on both iOS and Windows (Synctrain is the best client on iOS), create a books folder on both devices, and sync it, and you could open books from that folder in Paperback like normal and they would be synced. Or you could use Dropbox or iCloud Drive or Google Drive, since the iOS file picker lets you browse cloud services, and on Windows those are just synced folders.
@Jonathan Mosen
in reply to Quin
mastodon - Link to source

Matthew Campbell

How exciting! Congrats! Here's hoping Apple will get it together and review the iOS version, this sounds like it could be a great alternative to VoiceDream Reader which seems to be in it’s death spiral. Have you thought about implementing a syncing feature through a cloud service like Dropbox, OneDrive, iCloud, ETC?