StreamGrid – Open Source Multi-Stream Viewer (v1.2.0 Update)

Hey everyone,

I previously shared StreamGrid, an open-source app for watching and managing multiple streams in a customizable grid layout. I’ve been working on it since then, and I’m excited to share the latest v1.2.0 update with major performance improvements and cross-platform build support.

# ✨ Highlights

🎥 Multi-stream viewing – Watch multiple live streams or VODs at the same time.
🖱️ Customizable layouts – Drag, drop, resize streams on the fly.
💾 Save & share – Export/import your grid configurations.
🌐 Stream support – Works with M3U8, HLS, MP4, and more.
🚀 Cross-platform – Windows, macOS, and Linux (Electron + React + TypeScript).

# 🔥 What’s New in v1.2.0

Nearly instant startup.
Virtual rendering → smooth performance with 50+ streams.
Player pooling → reduced memory usage + faster stream switching.
Lazy-loaded chat components for better efficiency.
Cross-platform build support (Windows, macOS, Linux installers).
Local file stream support + improved error handling.
The ability to save "grids" and load them at any time allowing easy switching between layouts

The app is still 100% free and open-source. I’d love for the community to give the update a try, share feedback, or contribute!

👉 GitHub Repository

Thanks again to everyone who checked it out before!

https://redd.it/1n2gcnk
@r_opensource

Trying to get my dad to use open source i failed

My dad called me while he was with sellers trying to get him to buy Microsoft 365 for like 200 dollars , while i was telling him noo get libreoffice it’s open source and free the seller heard me he said to him it’s not Microsoft! and my dad is not tech savvy so he got confused they brought a female seller and also tried to convince him and now he thinks im the devil and a hacker and he doesn’t trust me xD

https://redd.it/1n2h3t2
@r_opensource

Looking for a FOSS PDF Drawing / Handwritten Note-taking app that is lightweight and cross platform between Windows, Android, and Linux.

# My Previous Setup

My previous workflow for hand notes consisted of three apps: Google Drive, Obsidian, and Excalidraw. On Windows, I used the Google Drive Client to create a virtual folder (via Windows Shell Namespace Extension), and I had Obsidian use that virtual folder as my vault. This way, all data was stored in Google's cloud while being accessible to Obsidian. I could then draw on PDFs via an infinite canvas for note-taking.

On Android, I used DriveSync to synchronize my Google Drive to a documents directory, and I then pointed my Obsidian vault to that documents directory. I could then draw on PDFs and create hand-written notes with Obsidian's Excalidraw plugin.

# My New Setup

Instead of Google Drive, I now have all of my data stored on my personal, self-hosted Nextcloud server. This includes my notes and several gigabytes worth of family photos. I no longer use Obsidian because Nextcloud's cloud markdown editor serves as a good substitute. Nextcloud also has its own document editor, spreadsheet editor, and powerpoint editor. I don't need Obsidian's document-linking feature in my workflow, so Obsidian has no advantages for me.

# The Problem

Nextcloud's implementation of Whiteboard (based on Excalidraw) is horrible for note-taking. It is slow (likely due to my server's slow CPU), but more importantly it does not support importing PDFs. So, I need an alternative. I want to do all my handwritten note-taking in my personal cloud ideally, but there don't seem to be any good Nextcloud integrations or apps for this. I don't want to use Obsidian for two reasons: I don't want to install an entire app just for Excalidraw, and I don't want to have to sync my Nextcloud files to a local directory due to the aforementioned several gigabytes worth of family photos blowing up my Android device's limited storage.

# What I've Tried

So far, I've tried Microsoft OneNote, Xournal++, Linwood Butterfly, Saber, and Okular.

OneNote works perfectly on Desktop, but OneNote doesn't support importing OneNote folders from local directories or the Nextcloud Document Provider on Android. It also doesn't support exporting to a local directory on Android. Everything must be done through Microsoft's cloud.

Xournal++ is amazing on desktop and does everything I want. I absolutely LOVE how imported PDF pages are given their own individual pages. This is great because it allows me to export the drawings I've made atop the imported PDF as regular PDF files. With Obsidian's Excalidraw, importing a PDF just dumped all the pages onto a single infinite canvas, so the only way to export the drawing was by exporting the entire, lengthy image. I also love that Xournal++ allows me to export its .xoj files directly to my Nextcloud virtual folder in Windows. However, the Android mobile app is no longer supported. It's last update was four years ago. Also, its ability to open files seems to be broken. When using the "open file" option, I cannot navigate to any directory in the file system because it only allows me to browse Google Drive and my Gallery. So, importing .xoj files on my Android device is impossible with the Xournal++ mobile app. This makes Xournal++ unusable to me.

Linwood Butterfly is pretty good because it allows me to import and export .bfly files to and from my Nextcloud virtual folder on Windows. The interface is intuitive and simple. However, I do not like how it uses an infinite canvas when importing a PDF for the same reason I dislike Excalidraw. Other than that, Butterfly works perfectly fine. However, it is a bit annoying on Android. It crashes whenever I try to import a .bfly file directly from my Android device's Nextcloud virtual folder. I can get around this by first copying my .bfly file from my Nextcloud virtual folder to a local folder and then importing the copied file. This gets tedious, but it works ok. Exporting a .bfly file from my Android to my Nextcloud virtual folder works

Looking for a FOSS PDF Drawing / Handwritten Note-taking app that is lightweight and cross platform between Windows, Android, and Linux.

# My Previous Setup

My previous workflow for hand notes consisted of three apps: Google Drive, Obsidian, and Excalidraw. On Windows, I used the Google Drive Client to create a virtual folder (via Windows Shell Namespace Extension), and I had Obsidian use that virtual folder as my vault. This way, all data was stored in Google's cloud while being accessible to Obsidian. I could then draw on PDFs via an infinite canvas for note-taking.

On Android, I used DriveSync to synchronize my Google Drive to a documents directory, and I then pointed my Obsidian vault to that documents directory. I could then draw on PDFs and create hand-written notes with Obsidian's Excalidraw plugin.

# My New Setup

Instead of Google Drive, I now have all of my data stored on my personal, self-hosted Nextcloud server. This includes my notes and several gigabytes worth of family photos. I no longer use Obsidian because Nextcloud's cloud markdown editor serves as a good substitute. Nextcloud also has its own document editor, spreadsheet editor, and powerpoint editor. I don't need Obsidian's document-linking feature in my workflow, so Obsidian has no advantages for me.

# The Problem

Nextcloud's implementation of Whiteboard (based on Excalidraw) is horrible for note-taking. It is slow (likely due to my server's slow CPU), but more importantly it does not support importing PDFs. So, I need an alternative. I want to do all my handwritten note-taking in my personal cloud ideally, but there don't seem to be any good Nextcloud integrations or apps for this. I don't want to use Obsidian for two reasons: I don't want to install an entire app just for Excalidraw, and I don't want to have to sync my Nextcloud files to a local directory due to the aforementioned several gigabytes worth of family photos blowing up my Android device's limited storage.

# What I've Tried

So far, I've tried Microsoft OneNote, Xournal++, Linwood Butterfly, Saber, and Okular.

OneNote works perfectly on Desktop, but OneNote doesn't support importing OneNote folders from local directories or the Nextcloud Document Provider on Android. It also doesn't support exporting to a local directory on Android. Everything must be done through Microsoft's cloud.

Xournal++ is amazing on desktop and does everything I want. I absolutely LOVE how imported PDF pages are given their own individual pages. This is great because it allows me to export the drawings I've made atop the imported PDF as regular PDF files. With Obsidian's Excalidraw, importing a PDF just dumped all the pages onto a single infinite canvas, so the only way to export the drawing was by exporting the entire, lengthy image. I also love that Xournal++ allows me to export its .xoj files directly to my Nextcloud virtual folder in Windows. However, the Android mobile app is no longer supported. It's last update was four years ago. Also, its ability to open files seems to be broken. When using the "open file" option, I cannot navigate to any directory in the file system because it only allows me to browse Google Drive and my Gallery. So, importing .xoj files on my Android device is impossible with the Xournal++ mobile app. This makes Xournal++ unusable to me.

Linwood Butterfly is pretty good because it allows me to import and export .bfly files to and from my Nextcloud virtual folder on Windows. The interface is intuitive and simple. However, I do not like how it uses an infinite canvas when importing a PDF for the same reason I dislike Excalidraw. Other than that, Butterfly works perfectly fine. However, it is a bit annoying on Android. It crashes whenever I try to import a .bfly file directly from my Android device's Nextcloud virtual folder. I can get around this by first copying my .bfly file from my Nextcloud virtual folder to a local folder and then importing the copied file. This gets tedious, but it works ok. Exporting a .bfly file from my Android to my Nextcloud virtual folder works

flawlessly.

Saber is so close to being good. It imports PDFs onto their own pages like Xournal++, and it has simple interface for both its Windows and Android applications. Importing and exporting to and from both Windows and Android virtual folders is possible. But, the deal-breaker for me is that the Windows version for whatever reason cannot import its own files. After exporting an .sba file on Windows, it becomes immediately impossible to ever import it back into the Windows Saber application.

Okular is way too bloated for my use-case, and its suite of drawing tools is limited compared to the others listed here. Okular also doesn't have an official Android version.

# What I Want

I'm looking for an app that:

* Is cross-platform between Windows, Android, and Linux
* Can directly import from and directly export to Windows' Shell Namespace Extension virtual folders and Android's SAF Document Provider virtual folders
* Free, preferably FOSS
* Lightweight and focused solely on providing a suite of hand drawing tools
* Ability to import PDFs to individual pages within its document format

https://redd.it/1n2i2cw
@r_opensource

I built an open-source image resizer that's 100% private (runs in your browser) and has a killer feature: you can set a target file size (e.g., "under 500 KB").
https://github.com/killcod3/image-resizer

https://redd.it/1n2mbls
@r_opensource

Any solo developers who open sourced their SaaS?

I am thinking to opensource the project a desktop app) I have been working on that has been really helpful for me during my studies. In the past I have built and shared few small opensource libraries, but they didn’t get much attention apart from few stars on GitHub. I believe it’s mostly because I didn’t talk much about it online after releasing it and only added 1-2 features since I didn’t need to extend it beyond what’s supposed to do.

However, this desktop app is different and I believe it could be bigger and I have more ideas to add to it with right community.

Anyone managed build and maintained an open source community as a solo developer? I eventually want to get some commercial benefit from the app I built, but I want to hit the right balance. Any good resources or videos on this you can share would be very much appreciated as well.

https://redd.it/1n2pvv4
@r_opensource

Open-source optimization engine for LLM agents 🧠
https://github.com/Handit-AI/handit.ai

https://redd.it/1n2oxbg
@r_opensource

Open source dilemma in the EU too: many see benefits, too few contribute
https://www.heise.de/en/news/Open-source-dilemma-in-the-EU-too-many-see-benefits-too-few-contribute-10624536.html

https://redd.it/1n2xr5k
@r_opensource

Made a tiny library to collect user feedback in your app

I built a small Android library called EchoBox, a lightweight way to collect in-app feedback from users (emoji + message) without needing a backend.

The library is Jetpack Compose for now, but I might add XML support later if people are interested. Contributions are very welcome!

GitHub: https://github.com/amitcodr/EchoBox

https://redd.it/1n32vf9
@r_opensource

Any opensource dictation(speech to text) application for windows? Alternative for the 'Dictate' option in Word?

I searched for one on the net, but most of them seem to be for Linux, Macos and Android.

I know of Sayboard and Whisper in Android.

I have seen whisper installation setups and have tried it, but havent't seen one that can directly dictate into other open apps.

Are there any applications like that which can directly enter the dictated text into a word or text file?
Like, one which can directy dictate into Word or Libreoffice Writer?

Thanks in advance.

https://redd.it/1n32tcy
@r_opensource

I've created a program to record streams

Hi all, I've just published a new software called TeleVedi. TeleVedi allows users to record streaming channels in real time or schedule automatic recordings for later.

It is:

\- open source

\- free of charge

\- equipped with a simple and intuitive interface

\- compatible with custom streams and playlists

\- available in English and Italian

\- currently compatible only with Windows 10 and 11

You can try it at https://github.com/Messina-Agata/TeleVedi



Waiting for your feedback :)

https://redd.it/1n35qzo
@r_opensource

Best alternative to Simplewall

After the creator of Simplewall left this.

What alternative for Windows Firewall + DNS do you recommend?

https://redd.it/1n35vbx
@r_opensource

Termix - Self hosted web-based SSH terminal, reverse tunnels, and file editing

GitHub Repo: [https://github.com/LukeGus/](https://github.com/LukeGus/Termix)

For the past couple of months, I have been working on my free self-hosted passion project, Termix.

Termix is an open-source, forever-free, self-hosted all-in-one server management platform. It provides a web-based solution for managing your servers and infrastructure through a single, intuitive interface. Termix offers SSH terminal access, SSH tunneling capabilities, and remote file editing, with many more tools to come.

Complete Feature List:

* **SSH Terminal Access** \- Full-featured terminal with split-screen support (up to 4 panels) and tab system
* **SSH Tunnel Management** \- Create and manage SSH tunnels with automatic reconnection and health monitoring
* **Remote File Editor** \- Edit files directly on remote servers with syntax highlighting, file management features (uploading, removing, renaming, deleting files)
* **SSH Host Manager** \- Save, organize, and manage your SSH connections with tags and folders
* **Server Stats** \- View CPU, memory, and HDD usage on any SSH server
* **User Authentication** \- Secure user management with admin controls and OIDC support with more auth types planned
* **Modern UI** \- Clean interface built with React, Tailwind CSS, and Shadcn

https://redd.it/1n35so1
@r_opensource

Drowning in pull requests from contributors with wildly different code quality

Maintaining an open source project and getting 10-15 PRs weekly from contributors ranging from complete beginners to senior engineers. The quality variance is insane.

Some PRs are production-ready, others introduce bugs that would crash the entire system. I spend more time reviewing and providing feedback than actually working on features. It's becoming unsustainable as a volunteer effort. The challenge is being educational without being discouraging. Want to help beginners learn but also need to maintain project quality. Can't just auto-reject low-quality PRs but can't merge everything either. Started using greptile to do initial screening and provide consistent feedback formatting. Helps catch obvious issues and gives me a starting point for more detailed reviews. Still working on finding the right balance between automation and human mentoring.

How do other maintainers handle this? What's your process for managing PR quality at scale while staying welcoming to new contributors?

https://redd.it/1n3c097
@r_opensource

Java Project to Contribute

Hey everyone
I’m Prakash, a software developer at Oracle with \~2 years of experience. I’m really interested in contributing to OpenSearch and would like to start with a project/module that’s both beginner-friendly and offers good learning opportunities.

I initially tried diving into the main OpenSearch repo, but found it a bit overwhelming 😅. Could you suggest some good sub-projects, plugins, or areas where new contributors usually start and can learn effectively?

My priority is to contribute to Java-based projects.

https://redd.it/1n3ckjy
@r_opensource

Sign the Petition
https://chng.it/rhbn9HG6Qm

https://redd.it/1n3h6l3
@r_opensource

We build and open-sourced a remote pair programming app
https://github.com/gethopp/hopp

https://redd.it/1n3j6ei
@r_opensource

Scenario-driven simulator for async backends (design before you code)

Hello!

I’d like to share AsyncFlow, an open-source scenario-driven simulator for asynchronous distributed backends.

With AsyncFlow you can do something like:

Declare a topology (users → client → load balancer → servers) → Run it → Inspect results (latency, queues, RAM, failures).

That’s the core idea: simulate your async system before it exists in production.

🔹 What AsyncFlow Does?

• Scenario-driven simulations → inject network RTT, jitter, outages, or resource caps.

• Event-loop aware servers → CPU blocks, RAM stays allocated, I/O yields like real async frameworks.

• Network edges → model latency distributions, spikes, or degraded links.

• Metrics out of the box → p95/p99 latency, throughput, ready queues, RAM, sockets under pressure.

• Event injection → simulate what-if dynamics (traffic spikes, server crashes, link degradation).

🔹 Who Is This For?

• Planners → test capacity and resilience trade-offs pre-deploy.

• Learners → see how async event loops behave under load.

• Educators → teach distributed systems with reproducible scenarios.

Repo: https://github.com/AsyncFlow-Sim/AsyncFlow

I’d love feedback on:

• Which scenarios (failures, network anomalies, overloads) you’d find most useful?

• What metrics would help you the most?

• Any OSS tips on docs and examples.

Thanks, happy to answer questions.


https://redd.it/1n3iknv
@r_opensource

Maintaining an Open Source Project in the Times of AI Coding

None of this text was written or reviewed by AI. All typos and mistakes are mine and mine alone.

After reviewing and merging dozens of PR's by external contributors who co-wrote them with AI (predominantly Claude), I thought I'd share my experiences, and speculate on the state of vibe coded projects.

tl;dr:

On one hand, I think writing and merging contributions to OSS got slower due to availability of AI tools. It is faster to get to some sorta-working, sorta-OK looking solution, but the review process, ironing out the details and bugs takes much longer than if the code had been written entirely without AI. I also think, there would be less overall frustration on both sides. On the other hand, I think without Claude we simply wouldn't have these contributions. The extreme speed to an initial pseudo-solution and the pseudo-addressing of review comments are addictive and are probably the only reason why people consider writing a contribution. So I guess a sort of win overall?

Now the longer version with some background. I am the dev of Serena MCP, where we use language servers to provide IDE-like tools to agents. In the last months, the popularity of the project exploded and we got tons of external contributions, mainly support for more languages. Serena is not a very complex project, and we made sure that adding support for a new language is not too hard. There is a detailed guideline on how to do that, and it can be done in a test-driven way.

Here is where external contributors working with Claude show the benefits and the downsides. Due to the instructions, Claude writes some tests and spits out initial support for a new language really quickly. But it will do anything to let the tests pass - including horrible levels of cheating. I have seen code where:

1. Tests are simply skipped if the asserts fail
2. Tests only testing trivialities, like isinstance(output, list) instead of doing anything useful
3. Using mocks instead of testing real implementations
4. If a problem appears, instead of fixing the configuration of the language server, Claude will write horrible hacks and workarounds to "solve" a non-existing problem. Tests pass, but the implementation is brittle, wrong and unnecessary

No human would ever write code this way. As you might imagine, the review process is often tenuous for both sides. When I comment on a hack, the PR authors were sometimes not even aware that it was present and couldn't explain why it was necessary. The PR in the end becomes a ton of commits (we always have to squash) and takes quite a lot of time to completion. As I said, without Claude it would probably be faster. But then again, without Claude it would probably not happen at all...

If you have made it this far, here some practical personal recommendations both for maintainers and for general users of AI for coding.

1. Make sure to include extremely detailed instructions on how tests should be written and that hacks and mocks have to be avoided. Shout at Claude if you must (that helps!).
2. Roll up your sleeves and put human effort on tests, maybe go through the effort of really writing them before the feature. Pretend it's 2022
3. Before starting with AI, think whether some simple copy-paste and minor adjustments will not also get you to an initial implementation faster. You will also feel more like you own the code
4. Know when to cut your losses. If you notice that you loose a lot of time with Claude, consider going back and doing some things on your own.
5. For maintainers - be aware of the typical cheating behavior of AI and be extremely suspicious of workarounds. Review the tests very thoroughly, more thoroughly than you'd have done a few years ago.

Finally, I don't even want to think about projects by vibe coders who are not seasoned programmers... After some weeks of development, it will probably be sandcastles with a foundation based on fantasy soap bubbles that will collapse with the first blow of the wind and can't be fixed.

Would love to

hear other experiences of OSS maintainers dealing with similar problems!

https://redd.it/1n3ms01
@r_opensource