I work on OpenSubs, a free, open source (AGPL-3.0) subtitle tool that runs entirely in the browser tab.

You drop in a video file, and Whisper transcribes it on your own machine, using transformers.js with WebGPU where available and WebAssembly otherwise. The model (40–250 MB) downloads once and is cached. There is no upload endpoint in the product, so the video has nowhere to go.

After that you can:

• fix lines by typing over them (click a timestamp to jump to that moment)
• translate into 20 languages with Chrome's built-in on-device translator
• pick one of 12 caption styles, including word-by-word highlighting
• export SRT / VTT / ASS, or burn the subtitles into an MP4 (libass compiled to WebAssembly, encoded with WebCodecs)

A few things I learned building it:

• Whisper hallucinates on silence and music ("Thanks for watching!", or the Japanese equivalent). A Silero VAD pass runs before Whisper, and a cleanup step drops the known stock phrases.
• Singing doesn't count as speech for the VAD, so a music video gets a "no speech found" warning. You can still force it.

Honest limits: it only takes video files, not audio-only files. Cue timings can't be edited yet. Builds are release candidates. Everything that runs locally is free with no account; the only paid part is optional cloud translation on our backend (US$5 for 1000 credits), and you can bring your own Claude / OpenAI / DeepL key instead.

Site: https://opensubs.app/ Code: https://github.com/open-subs/opensubs

Feedback welcome, especially on languages where the transcription goes wrong.

Google's September 2026 Pixel Update Bulletin contains patches beyond what's in that month's regular Android Security Bulletin. According to GrapheneOS, some of those extra patches touch standard Android platform code, the kind that runs on non-Pixel devices, not just Pixel-branded hardware.

None of that platform-level code has reached the regular monthly bulletin or the private preview patches other manufacturers typically draw from to get their own patches ready.

And at this rate, these won't reach non-Pixel OEMs at all until Android 17 QPR2 ships later this year in December.

The project is characterizing this as Google "gatekeeping security patches to the standard Android platform code from Android OEMs." The complaints

GrapheneOS says Android 17 QPR1 shipped new developer APIs that never made it into AOSP. This is something they claim hasn't happened since Android's Honeycomb days.

Google's API diff report backs this up. Comparing Android 17 to QPR1 shows one new package, android.hardware.hid, plus changes across sixteen others, including android.media, android.os, android.provider, android.telecom, and android.view.

GrapheneOS has ported its code to QPR1 before Google even released it, but still doesn't have permission to ship that work. For now, the project is backporting Pixel firmware, kernel drivers, userspace drivers, and HALs from QPR1 onto Android 17 instead.

On top of all that, there's a compliance issue that seems to be recurring.

Google was slow to comply with a GPL source request. GrapheneOS requested sources for a build (CD1A.260905.001.A1) on September 1, and access only came through more than two weeks later.

Why this is worrying

None of these three issues is catastrophic by itself. A three-month patch delay, a paused API rollout, a two-week wait on source code—each is the kind of thing that could pass as a one-off.

Taken together, however, they point to a recurring theme. Google is holding security fixes back from the wider Android ecosystem, withholding new APIs from AOSP for the first time in over a decade, and slow-walking GPL compliance it's required to meet.

Don't even get me started on what they are doing to the Android app ecosystem.

Google is on track to require every Android app developer, whether on the Play Store, F-Droid, or anywhere else, to register with them. Come 2027, that means handing over legal identification and signing key evidence before an app can run on any certified device.

Sideloading an unverified app following this would mean enabling developer settings, waiting through a mandatory 24-hour cooldown, and clicking past several warning screens (classic scare tactics, btw).

GrapheneOS is one of dozens of organizations that signed onto the Keep Android Open campaign opposing this, alongside F-Droid, the Electronic Frontier Foundation, and the Free Software Foundation.

If you ask me, this Big Tech company is doing what's regrettably natural for it, clamping down open access to things so that its competition cannot benefit.

Title.

I have looked for SO long, I've tried every icon pack on f-droid, I've looked through the play store multiple times (either they were ugly or collected all of my precious data 🥹), I've even tried to make my own icon pack with renkin but it barely works (I'm not interested in fixing it) SO I just need some icon packs to look through

also I'm pretty picky, and I want either funny/kinda cursed or something that isn't flat design and monochrome

thank guy I love drew gooden

Android 17 QPR1 is the first release since Android Honeycomb (3.x) adding new APIs for app developers without a release to the Android Open Source Project. The new APIs are currently exclusive to the Pixel OS and aren't available to other Android OEMs.

https://developer.android.com/sdk/api_diff/37.1/changes

Non-Google Android OEMs and AOSP-based projects can ship yearly and QPR2 releases. There are also security backports to those releases. Security preview access is needed to ship patches without months of delay. We've had that since before our Motorola partnership via another OEM.

We already ported our code to Android 17 QPR1 since before it was released on September 15th but don't have permission to release it yet. We're working on backporting Pixel firmware, kernel drivers, userspace drivers and HALs from Android 17 QPR1 to Android 17 for now instead.

Pixel Update Bulletin for September 2026 has additional patches to standard Android platform components used by non-Pixel devices. These patches are relevant to non-Pixel devices but haven't been made available via the September 2026 Android Security Bulletin or preview patches.

Google should not be gatekeeping security patches to the standard Android platform code from Android OEMs but that's what they've started doing. Other OEMs will get these patches in December 2026 via Android 17 QPR2. We can ship them early by reverse engineering the code instead.

It would be interesting to know if Google's legal team is aware they're giving Pixels months of early access to new Android features and bug fixes including certain important security patches. Pixels being given this competitive edge over Google's OEM partners is very dubious.

Google is also once again failing to comply with our GPL source requests for weeks. We requested CD1A.260905.001.A1 sources on September 1st and were only provided access today. The kernel build IDs are the same for 17 QPR1 Beta 9 and 17 QPR1 so we have that already at least.

We could have shipped an Android 17 QPR1 update already since our port was completed. Instead, we have to deal with ongoing pain until Android 17 QPR2 is released in December 2026. This will not be an issue for Motorola since we'll have official firmware and driver code provided.

It will be far easier for us to support upcoming Motorola devices than Pixels. Qualcomm will hopefully expand MTE support beyond the highest end flagship SoC soon so we can expand to more than flagships. Pixel 11 didn't remove it from hardware, only firmware, so that's good news.

Running local LLMs on consumer AMD GPUs often comes with two pain points: configuring backend overrides (HSA_OVERRIDE_GFX_VERSION) and figuring out whether Vulkan or HIP/ROCm gives better token throughput on your setup.

I created two open-source tools to solve this:

1. ROCmFix Repo: github.com/xanpavle/rocmfix

Zero-dependency Python script runnable via single-line terminal command.

Detects GPU PCI IDs directly (Windows Registry / Linux lspci) to map the correct HSA_OVERRIDE_GFX_VERSION.

Automatically updates shell profiles (PowerShell, CMD, Bash, Zsh, Fish) and includes an undo command and driver diagnostic runner (rocmfix doctor).

2. InferBench Repo: github.com/xanpavle/inferbench

Automated benchmarking tool for local LLM servers (Ollama, LM Studio).

Rigorously compares Vulkan vs HIP/ROCm performance with warm-up passes, median of N runs, and cold-start VRAM unloading between test cycles.

Both projects are MIT-licensed. Check them out and let me know your benchmarking results!

cross-posted from: https://programming.dev/post/56782117

Supply-chain through social engineering

Open source printer with refillable ink, no bullshit, accept rolls of papers.

Finally an open source printer ! I dealt with so much stupidity with printers. it warm my heart.

I do have too much joy to practice mandatory critical thinking about the release of a new project. Lets talk about it? i am not the only one to be happy about it, i have no info about the price though :/

The latest update to Calibre, the popular open-source ebook manager, features an AI-powered writing game.

Calibre v9.15’s “Create Your Own Adventure” mode, per the release notes, lets you build a fictional world and hand control of it to an AI who will manage it in real time based on your responses. A bit like old text-based RPGs, I guess.

The changelog notes, with obvious understatement, that the feature is “a little different” from Calibre’s usual fare, adding it began as a way for developers to test the tool’s AI backend before developing into “something fun to use”. Calibre Create your own adventure AI feature wizard Setting up your AI world in Calibre v9.15

Like all AI features in Calibre, it’s completely optional as it relies on you use adding an API key for a compatible LLM. To try it, go to Preferences > Toolbars and add the “Create your own Adventure” button to a toolbar. Click this to open a setup wizard.

This isn’t the first puzzling AI feature the Calibre has gained. Last month it added hooks to generate ebook covers using AI prompts and preset styles (a noir Wuthering Heights; a pulp sci-fi Barbara Cartland novel, and so on).

Calibre users are, largely, people who like reading books – most with a long backlog of titles they’re not getting through1. Folks who like using LLMs to generate fiction likely have no need for an ebook manager, and vice versa.

I do wonder if Calibre ought to focus on adding AI features that help, rather than distract: letting users sort ebooks, fix metadata or generate “playlists” of what read next based on a users’ library and tastes.

Generating slop covers, summarising books and, now, generating fiction feels2 at odds with its role. Other changes in Calibre 9.15

This update also ships with improvements that don’t cause a collective head scratch.

The Highlights panel in the ebook viewer lets you select multiple highlights and right-click them to change the style, and if a highlight has notes, you can see that note in a hover tooltip when you mouse over it (this behaviour can be disabled).

More customisations are present. A button can be added to the sidebar to change between saved window sizes, and in the cover grid you can now choose which corner of book emblems appear via Preferences > Look & Feel > Cover grid > Layout.

Plenty of bugs were fixed, including issues with Read Aloud stopping at large images, a spellcheck quirk that overlooked words with a single apostrophe and Calibre not starting for users in Alberta Canada after its recent timezone change. Where to get Calibre

Calibre is free, open-source software for Windows, macOS and Linux.

Download Calibre from the official website – on Linux, Calibre is installed and updated via a command-line script that will download a binary, unpack it to a system folder and create the various .desktop shortcuts needed to find and open it in app menus, etc.

Alternatively, you can get Calibre on Flathub, although it is an unverified build.

https://gitlab.com/christosangel/corsaro

corsaro is a script written in Bash that allows the user to search for torrents in the terminal .

The functions and the data are taken from https://piratebay.live/

Opening the script, the user through the main menu, can:

• Search for a torrent

• Search for a torrent in a category

• Search for a torrent in multiple categories

• Check out top 100 torrents in a category

• Browse torrents in a category

• Browse the most recent torrents

• Repeat previous torrent search

• Browse torrent search history and repeat a search using a stored key word

• Configure the parameters of the script

https://f-droid.org/packages/com.sidhant.huesort/

I love hue

starts easy, quickly gets difficult. Don't forget to turn off night light, red shift &c to play

Beware the Computerworld's article 🙄 that's about NATO Office 🙄🙄

Nextcloud wants to close the gap with Microsoft Office with the introduction of a new Euro-Office desktop client for Windows, macOS, and Linux — one of numerous updates to its Nextcloud Hub suite.

Nextcloud has developed a desktop app for its Euro-Office-based productivity applications, filling a gap in the software suite that it pitches as a sovereign alternative to Microsoft Office.

Nextcloud, one of the backers of the Euro-Office initiative, integrated the software into its Nextcloud Office suite in June. Until now, it’s been accessible via a web browser or mobile app.

The new desktop client lets users edit documents, spreadsheets, and presentations locally, and sync files online for collaborative editing. It will be available for Linux, macOS, and Windows in the coming weeks.

“For organizations and individuals who prefer a traditional desktop experience, this closes one of the last gaps between Euro-Office and Microsoft Office, without handing control of their data to a non-European vendor,” Nextcloud said in a press release.

Rather than a standalone application suite, Euro-Office is designed as an integration component that’s embedded into productivity applications, such as Nextcloud Hub. Euro-Office handles document editing, while other elements such as storage and permissions managed by the wider platform.

The open-source project, which is built on OnlyOffice’s codebase, is backed by Nextcloud as well as Eurostack, Ionos, Office.eu, Proton, XWiki, and others. More than 40 individuals have contributed code to Euro-Office, according to Nextcloud, with around 15% of contributions from outside of the organizations directly involved in project.

Euro-Office supports OOXML file formats for compatibility with Microsoft Office, as well as OpenDocument Format (ODF) files, including ODT, ODS and ODP. The Document Foundation, which is responsible for development of LibreOffice, another open-source productivity suite, has previously questioned the Euro-Office initiative’s preference for Microsoft file formats over ODF.

Organizations using Euro-Office include Germany’s Deutsche Telekom Security and Destek, a Turkish IT consultancy, according to Nextcloud. Other customers in countries such as France, Switzerland, and Philippines are currently evaluating the software.

But it’s still early days for Euro-Office. Open-source software initiatives often face a challenge in gaining traction among a broad enterprise audience, said Martha Bennett, VP and principal analyst at Forrester.

“Euro-Office hasn’t really proven itself at the kind of scale that large enterprises are looking for. That’s of course not unique to Euro-Office; all potential open-source alternatives face the same issue,” she said.

“Obviously, it’s a bit of a chicken-and-egg situation — without somebody taking the leap, it’s not possible to provide that proof.” Nextcloud Hub Summer 26 updates

Among the other announcements with the Summer 26 updates are changes to management of teams within Nextcloud Hub. An updated Nextcloud Teams app, which is now available in the app menu, provides users with an overview of the teams they belong to. Each team features a folder with shared resources such as files and tasks, which Nextcloud notes is similar to Microsoft’s approach with Teams and SharePoint.

Nextcloud now also enables the creation of team-owned files, starting with Nextcloud Collectives and Deck, with additional apps in future. The aim is to ensure that all team members can access a file, even if the creator leaves the team workspace. For admins, there are new tools to view and manage team folders, including configuring storage quotas.

Among the other announcements are updates to Nextcloud Text — a lightweight text editor available across several apps — with version comparisons, footnotes, and references available throughout a document, and inline comments. Nextcloud Tables, a database app, now has relational columns that link data in cells across multiple tables, and the ability to import and export table structure — such as columns, views, and table-based apps — without changing data.

With Nextcloud’s AI Assistant, users can now upload an image and ask questions about it or translate text, while the chatbot will also recall information across conversations.

UI improvements include a revamped search experience that displays results from across Nextcloud Hub files, messages, and events, with filter options such as date range and people. App menu icons have been updated to make apps easier to recognize, and new animations throughout the interface make navigation and interactions “feel more fluid and easier to follow,” said Nextcloud.

cross-posted from: https://discuss.tchncs.de/post/67770797

Hey everyone!

I have been building and maintaining my own infrastructure around self-hosted Linux services for months. Eventually I got annoyed with editing and deploying everything manually and having difficulties sharing what I'm doing with others. So I solved both of these problems with Infrastructure as Code.

Everyone is free to study this repository. Ever wanted to see IaC in practice? Curious how to self-host certain services, automate website deployment from Git to server, make TLS SNI-based router or automate container provisioning? Then this repository is for you.

Want to know how much CPU, RAM or other resources a certain service has? You can see it yourself on my own public dashboard.

This is the first time I share my infrastructure with anyone. I'd be more than happy to discuss it, receive feedback and answer question. In the meantime, I'm writing an article to complement the code with architecture, decisions and thought behind it.

About a decade ago, I tried to index the web via the Dewey Decimal System. I had a site laid out similar to Google, where you could browse sites continuously starting from a given call, but the DDS is proprietary, and those people hate anyone who uses their IP without a license. You can Google everyone they've shutdown -- places that weren't even libraries -- for using anything similar to the DDS. I reached out to the group that manages the DDS, and was taken offline before my project even started.

With all the corporate BS lately, and people looking for alternative options, I thought I'd take a search engine old school, and we'd index like Usenet. Except with an XXX.XXX.XXX format.

Problem is, now I've got to find categories, sub categories, and local categories. Categories for animals, chat rooms, cars, construction, wrestling, etc.

I need a brilliantly created categorical system with a TLD.MLD.LLD (top level subject domain, mid level domain, and local level domain) format.

I was going to let site owners Disney Land it, by letting them create an index.json file in the root of their site, choosing for themselves where they wanted to appear in the index. Problem is, as trends change and systems become more exact, site owners would have to update their index. Which could become cumbersome.

Further, if we divided into academic disciplines, there would be things like gaming and model construction which wouldn't fit into those.

It would help if there were an already established categorization system, so that site owners didn't have choice anxiety. And then, as the site grew, site owners could recategorize themselves with user feedback.

Could you guys suggest a few categorization systems (can't be DDS), either of your own making or other sources?

I'm already looking at Usenet: https://www.livinginternet.com/u/uw_hier.htm

Edit: Current UX: https://imgur.com/a/sURm4s1

• Haven't chosen a name yet.
• Adult is probably going to be dropped, and when the site owner chooses a category, they can put it in the ADT TLD.

Hi all,

We're looking to get off Rackspace's email as the Spam filtering is pure garbage. We're going to avoid the Googles and MS's of the world, and I'm wondering if this community has suggestions for mail providers that run FOSS services (and are public about that via support), and/or are privacy focused. Note: Proton is a no-go due to having to run an IMAP middle-man service.

Thanks!

It's just a proof of concept.

https://concord.polarisocial.xyz/

A complaint-driven, consensus-based forge: development is driven by real complaints, ranked by pairwise Glicko-2 comparisons, decided through quorum consensus, executed on a phase-gated kanban board — and discovered through a first-class search surface (tags, quality, maintenance health, languages).

Governance is collective by default: maintainers exist, but direction is decided by collaborators. See docs/concord-spec.md.

cross-posted from: https://aussie.zone/post/36604349

This appears to be a minor bugfix release for the recent 12.0 major.

You can find more details about and discuss this release on the official forums.

As always, please ensure you stop your Jellyfin server and take a full backup of your metadata/configs before upgrading!

i'm searching for an audio player that is suitable for radio productions. the radio format has some talking and some music playing. the playlist of songs is prearranged, so normally no shuffling of music during the radio show.

so ideally the software would have one playlist window and one big button "play next" that plays the next song, and once the song is over it pauses. or maybe (but this is optional) it can be configured to play the next two songs, because two songs are also possible between the talking.

another way to describe it:

• you should be able to create a playlist
• you should be able to set pauses/breakpoints after one or multiple songs

nice to have:

• you can see how long the song still takes
• maybe show audio histogram or spectrogram so you now when you can start talking/have to stop talking

mixxx is quite close, but i can not find a workflow that supports the scenario above.

amarok has a stop after that song feature, but you can only set one of those stop points, i would need multiple, the idea is to setup the playlist before, but during the show, you don't want to prepare the next song or stuff, you just want to press a key, or the big button.

maybe a bash script with mpv calls and read in between would be enough, but it would be nice if a normal person could also use it, and you don't need to edit a bash script to create the playlist. and of course i've only described the happy path above, maybe it would also be nice to jump back in case you made a mistake, which would be hard with the bash script.

thanks for your input!

Mozilla today published their Firefox 156.0 release binaries ahead of the official Tuesday announcement.

With Firefox 156 they are promoting their built-in PDF viewer as now starting up to 45% faster. Also on the performance side is improved memory and CPU usage when Firefox is displaying large JPEG images that in turn are scaled down to fit on a web page.

There are also a number of fixes in Firefox 156 like for high-sample-rate FLAC audio in MP4 failing on some websites, fixing various platform-specific issues, and other minor adjustments.

Within the Firefox 156 nightly builds was also enabling Vulkan Video decoding for newer NVIDIA GPUs on recent versions of the NVIDIA driver, but sadly that didn't make it for the Firefox 156 release with it still being gated to nightly builds.

This isn't the most exciting Firefox update for Linux users in recent history, but if wanting this latest update you can grab Firefox 156 right now via ftp.mozilla.org.