Help us pick an open-source product to build in 12 months - tell us your real pain points

Small CS team at the university with a full year for a school project (which needs to be released as open source) wants to build and ship one useful, privacy-respecting open-source product. We’ll work in public, maintain it after 1.0, and we’re looking for your real, recurring pain to solve.

https://redd.it/1niihr8
@r_opensource

I'm a product designer and i want to collaborate with other devs

Hello everyone!

I've been getting into product design for the past two years. I see a lot of devs with cool projects but it's too much work to design, develope, and then market and test their products properly without funding. so I want to collaborate on projects and hopefully offer some insight and help you

I’m especially interested in teaming up on something valuable, and maybe even shaping it into something we could grow or monetize down the road. If anyone’s working on a project that could use some design, I’d be excited to chat and see where it goes.

feel free to DM me or email me at yaserbustati@gmail.com if you prefer
thanks!

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

WhoAmI.tech has a new update! Hopefully it helps some of you!!

Hi everyone,

So a while back I made a post about WhoAmI, a free application that I built for people in tech to build a web presence. It's a sort of mix between LinkTree and LinkedIn !

Long story short, so many people signed up and were really positive in the comments and it meant so much to me that I had to keep working on it.

But feelings aside, now there's a new update with some cool features I think you'll enjoy:

- RSS Feed

Every profile now has an RSS feed that users can subscribe to. You can see it on my profile at https://whoami.tech/cfds on the posts section.

- Newsletter feature

Every public post you create now has a newsletter subnoscription form as you can see here https://whoami.tech/cfds/posts/working-on-whoami-s-profile-page (you can take the chance to subscribe to my newsletter if you want :P )

When the user subscribes, they receive a confirmation email and once confirmed, the subnoscription is active. The profile owner you subscribed too can see their number of subnoscriptions on the dashboard.

Now, once they create a post they have a checkbox to send to the newsletter or not, and if they do, everyone in it receives the new post per email !!

Of course there's a link available to unsubscribee as well :).
____

Again the app is free for everyone to USE, I might at some point implement a Buy Me a Coffee button for whoever wants to help cover the cost of the VPS and Domain (like $5) :D, but hopefully it helps some of you !!! I've been so excited and thank you for making me feel special and supporting this project :D.

https://github.com/s1lvax/whoami

https://whoami.tech

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

(de)centralized in trust or black box

Most AI demos look slick until you try to run them for real. When it’s your own data and your own infrastructure, verifying what an AI agent did becomes crucial. We’ve been exploring an open‑source stack that connects AI models with a decentralised data layer. The idea is: you send in an intent and get a result you can prove end‑to‑end. Everything runs on open‑source components (including the models when possible), and data stays yours, encrypted.

On paper that sounds neat; in practice, it raises some tricky questions. How do you guarantee privacy and still trust the results? Where do you draw the line between building a framework that others can compose and a finished tool someone can pick up and use? And what does verifiability even look like in the context of LLM agents?

I’m curious how others in the open‑source community are thinking about this. Have you tried combining AI with sovereign or decentralised systems? Do you lean towards building infrastructure or user‑facing apps? How do you approach auditing or proving what an AI agent has done? Sharing our experience here to start a conversation and see what patterns or pitfalls others have found.

https://redd.it/1niop03
@r_opensource

I built an open source alternative to piano learning tools
https://github.com/Tbence132545/Melodigram

https://redd.it/1ninpt8
@r_opensource

Vendure in 100 seconds
https://www.youtube.com/watch?v=atu89yudOxA

https://redd.it/1niqs81
@r_opensource

An open source caffeine-quitting app
https://github.com/drg101/decaf

https://redd.it/1niszm5
@r_opensource

Major Update: Tally.so n8n Community Node v1.2.0 Released

🚀 Just released a major update to the Tally.so n8n community node!

Highlights:
- Copy entire forms across different Tally accounts in one workflow (Get Form → Create Form).
- New field operations (add, update, delete, rollback).
- Safety features like Dry-Run preview + backup JSON for safer automation.

If you’re using n8n + Tally.so for workflows, this should save a lot of time. Feedback welcome!


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

Offline EPC/SEPA QR Code Generator (static web app)

I built a tiny utility for anyone who needs EPC/SEPA payment QR codes without trusting an online service. Everything lives from a single HTML file—open index.html or try the GitHub Pages build—and it runs entirely offline in your browser. You get live IBAN validation with readable spacing, mutually exclusive payment reference or RF structured reference fields, optional purpose code/BIC/note inputs, and a running byte counter to keep you under the 331-byte EPC payload cap. Once you’re happy with the payload you can export a crisp QR as PNG, JPG, or SVG. There’s also a dark/light theme, tooltips, example data, and built-in localization for all EU-SEPA countries.
Demo: https://quasistatic-setup.github.io/EPC-QR-Code-Offline-Generator/
Source: https://github.com/quasistatic-setup/EPC-QR-Code-Offline-Generator
I’d love feedback—edge cases to cover, translation contributions, or UX tweaks you’d like to see.

https://redd.it/1nit4ri
@r_opensource

Are all Tuta & Proton apps open source?
https://x.com/TutaPrivacy/status/1967240420108157079

https://redd.it/1nimdk4
@r_opensource

Advice: Etiquette for supporting a 'demanding' person in an open-source project

There's a piece of open-source software I use as a hobby, which has a relatively small community of fairly dedicated users. This software is written in C++ and has an embedded JavaScript interpreter, which allows users to write JavaScript mods/noscripts to provide additional functionality without modifying the C++ source.

I've written multiple mods for it in JavaScript and have shared my mods with the community. There's another user who has talked to me repeatedly with issue reports & feature requests for my mods, which is fine. However, one thing he requested some time ago is basically a whole functional NNTP client (newsgroup reader)) in JavaScript. Mind you, it's text-based, so it doesn't have a GUI. I've actually completed a large bulk of it; I think one major thing remaining is to have it clean up message text, which may have text in quoted printable format.

I think the reason he has asked me to write this for him is, as he has said, he "can't be bothered" to really learn JavaScript; it sounds like he's unwilling to learn JavaScript and wants others to do a lot of the work for him in creating these JavaScript mods he wants. It sounds like he has done programming in the past, so I don't think he's entirely unfamiliar with software development.

Normally, the JavaScript mods I write for this project are things I also use. However, I don't plan to use this newsgroup reader myself. While I like developing software, for a hobby project, I'm not quite as interested in developing something I'm not going to use personally. This would all be for him. Sometimes I've thought about telling him he can take what I have and finish it himself - I think he'd be in a good position to do that; Since he's the one who will be using it, he will be able to identify any issues quickly, and then he can fix them. Is that reasonable?

Another reason I'd like to just give it to him is because he can also sometimes be a bit condescending in the way he talks to people like me for support. I also feel like he can be a bit demanding. He frequently requests updates, which can feel tiring (though many of which are bugs he has identified, which is good). In the past 3-4 years or so, I'd guess about 95% of the change requests for my JavaScript mods for this project have been from him. I don't really feel like supporting something that I'm not even going to be using.

https://redd.it/1niymp7
@r_opensource

Balanced Ternary Abacus | Heisanban

Inspired by the Japanese Soroban, 平三盤 (Hei-San-Ban) is a computational abacus (physical or digital) designed for calculations in the balanced ternary number system.

Unlike traditional systems, it uses digits {-1, 0, +1}, enabling a symmetric representation of positive and negative numbers.

The name captures the project’s essence:

平 (Hei): system balance

三 (San): numeric base three (ternary)

盤 (Ban): the board/apparatus (abacus)

This repository hosts an interactive implementation of the Hei-San-Ban, serving as **an educational tool** and a practical exploration of balanced base-3 computation.

Use it online 👉 [https://robsoncassiano.software/tools/heisanban](https://robsoncassiano.software/tools/heisanban)

Repository 👉 [https://github.com/RandintN/abaco-ternario-balanceado](https://github.com/RandintN/abaco-ternario-balanceado)

Features

* Responsive, interactive UI (top/bottom beads touch the center bar)
* Decimal total and MathJax-rendered notation
* Bilingual content (PT/EN) with a toggle button
* Soroban-inspired tips and foldable tutorial/add/subtract sections
* PWA with offline support after the first visit

Live Long and Prosper 🖖🏻

https://redd.it/1nirml4
@r_opensource

Alternative vector graphics programs (not inkscape)

Hi all, I'm interested in learning vector graphics, but at a more basic level. I don't need all the power and complexity of Inkscape, and I've tried to learn it a few times without success. For reference, I'm a big user of paint.net (I know, free as in beer, not software, but it's what I'm used to) over gimp for most tasks, for the same reason. I don't need a lot of power out of my image editors, but easy and fast are key.

I get that Inkscape is supposed to be a free and easy alternative to Illustrator, but it feels difficult to me. I shouldn't have to Google a tutorial to draw an arc, but, as a beginner, I needed to last week. It feels like this is the case for every basic operation.

I'm going to be using this tool to make woodworking templates and laser cutting paths, so it can be very basic, but, especially for the laser cutting, it does need to be vector, and because of how important quality snapping is for the woodworking templates, vector makes sense there as well.

Because of this use case, I don't even need color support (though it would be very nice to have). I just need to be able to draw shapes, snap them together, sketch out curves and the like. All things Inkscape can do, but not things that are, in my opinion, intuitive to do in such a powerful tool.


https://redd.it/1nj2qt6
@r_opensource

Got blocked from GitHub org after delivering $200 bounty solution

Five years ago a maintainer put up a $200 bounty to add Windows UI Automation support for Dragonfly/Caster's accessibility. I took it and shipped in 2 days. Then I had to rebuild the work twice because the guidance kept changing. When I finally delivered what they originally asked for, I was blocked from the org for "AI usage", despite getting explicit approval to use AI responsibly and there being no written policy.

Quick background so you know where I'm coming from. I've been doing open source for about 4 months now. Usually maintainers are pretty reasonable. I've gotten write access to a few repos, contributed to some bigger projects. This is the first time I've run into anything this ridiculous.

Let me explain briefly:

The issue was clear: implement a UIA backend for dragonfly's accessibility API. The existing system used IA2 which had limited Windows support. They wanted something better.

I asked to work on it. Maintainer says "Please be responsible with AI if it is used" and gives me the green light. For context, I was using GitHub Copilot for code reviews, boilerplate, and drafting PR messages. Nothing more. All the actual implementation and technical research across their three repos was me.

First attempt: I proposed implementing it in dtactions (their utility library). Built a complete cross-platform accessibility adapter with Windows UIA support. Full test suite. Clean API. Ready to go.

Then the maintainer changes his mind: "One change from the original op is not to put it into dtactions. dtactions never really took off."

Second attempt: New guidance was "Dragonfly already has an accessibility controller which abstracts from the OS and grammars, which we already have a caster grammar. You could simply add it to a new controller there and everything would work."

So I pivot completely. Build a Caster grammar that calls dragonfly's existing accessibility controller. Different architecture, different approach. 214 lines of grammar code plus tests.

Third attempt: Submit the PR. Maintainer feedback: "The reference issue is describing about implementing another backend for the Dragonfly accessibility API."

Wait, what? Now they want a backend again? After explicitly telling me to use the existing controller?

I pivot AGAIN. Build a complete UIA backend. 483 lines. Thread-safe COM operations. Full dragonfly interface compliance. Real-world testing with Windows applications.

Then things went sideways. First red flag: maintainer accuses me of being an "AI bot" and closes the PR without review. This is after explicitly approving AI use and watching me implement three different approaches based on their changing guidance.

I push back professionally; called him out. Point out he approved AI usage. He reopens.

Then the dragonfly maintainer shows up demanding proof of "working with Caster, a Dragonfly SR engine and a specific UIA-enabled application." That's asking for a full speech recognition demo. For a backend implementation.

I provide a screen recording showing the UIA backend detecting real Windows applications, reading their properties, interfacing with the accessibility APIs. Exactly what a backend should do.

"You have not provided what I asked for. And that's because you can't."

Translation: they want me to set up an entire speech recognition pipeline. Microphone, speech engine, grammar system, the works.

When I question this impossible requirement:

"Yes, that is what I am requesting. Neither of us are comfortable with your posts or AI usage. This was not an issue four years ago when the issue in question was opened and, clearly, we need to have a written policy about it."

Let's break this down:

They admit they want an impossible demo
They're "uncomfortable" with AI usage they explicitly approved
They admit they have no written policy
They're enforcing rules that didn't exist

Then immediately blocked me from the entire organization.

I built three different implementations following their changing guidance. Each time they

moved the goalposts. When I finally delivered exactly what they originally asked for, they invented new requirements and blocked me for following their own AI usage policy.

If maintainers can retroactively enforce non-existent policies and change requirements mid-implementation, no contributor is safe.

Evidence:

[Issue #814](https://github.com/dictation-toolbox/Caster/issues/814) \- the original request
PR #950 \- my final implementation
Issue and PR comments showing the complete timeline
Blocking message \- proof of org ban

Is this acceptable maintainer behavior? Should contributors have recourse when maintainers act in bad faith? How do we prevent this from becoming the norm?

Because if this is where open source is heading, we have a serious problem.

https://redd.it/1nj68kq
@r_opensource

What is PyBotchi and how does it work?
https://github.com/amadolid/pybotchi

https://redd.it/1nj7irm
@r_opensource

Paywalls, licence switches… where’s the line for open source?

In the past two years a number of “open source” companies have quietly shifted from permissive licences to “non-compete” or pay-walled models. MariaDB introduced the Business Source Licence (BSL) in 2016; MongoDB, Confluent and Redis Labs followed; and HashiCorp switched Terraform to a non-compete licence. The justification is almost always the same: as these companies grow, the financial upside of being fully open diminishes, so they try to cut off “freeloaders” and capture more value. But the backlash is real: users and competitors fork projects and publish manifestos warning that licence switches create legal risk.

Red Hat’s decision to remove public access to RHEL source code has hit a similar nerve. SUSE’s Dr. Thomas Di Giacomo notes that RHEL exists only because of upstream projects like the Linux kernel, and Red Hat’s move has caused “significant concern within the open source community.” He argues that the freedom to access, modify and distribute software should remain open to all.

At the same time, many maintainers who make the code that powers our systems aren’t being paid. A 2024 Tidelift report found that 60 % of maintainers remain unpaid. The same report called this a “tragedy of the commons”: companies use free software without contributing code or funding. Burnout is inevitable; one developer with nearly three-quarters of a million downloads says he receives “no money at all.” Advocacy groups now propose that companies pay maintainers directly, for example; the OSS Pledge suggests $2 000 per developer per year.

So where’s the ethical line? At what point does gating features or switching licences move from sustainable funding to a betrayal of open-source values? Should we accept freemium models as a way to pay maintainers, or do they undermine the freedom that made Linux and FOSS so powerful? Curious how others here see it.

https://redd.it/1nj8oot
@r_opensource

Idea: logical fallacy detector

I don't build software but have an idea I think would help people (including me) - so throwing the idea out there for anyone interested:

TLDR: video logical fallacy detector

Problem:
Regardless of your political views, I think it's fair to say most Internet is an echo chamber for what you already think and many get their information for 30 second video clips.

Idea: (rough idea)
Browser plug in? that shows a small icon whenever a logical fallacy is used - straw man argument, appeal to authority, ad hominem, etc. ideally could be used when browsing YouTube or any other social media.
Small icon ideally would be clickable to give more info on why it's a fallacy, optionally fact checker as well.

I would gladly pay for a subnoscription to this. I have found similar but they are text only, and I believe a big misinformation issue is the short videos people watch.

Brainstormed the idea with gpt to get an elevator pitch:
“Think of this like a fact-checker for arguments. It’s a browser add-on that watches YouTube / X / Facebook/ etc with you and pops up a small symbol whenever someone is using a trick in reasoning — like attacking the person instead of the idea, pretending there are only two choices, or jumping to conclusions without evidence. You’d just click the symbol to see a quick, plain-language explanation of what happened. To build it, you’d tap into video captions (or speech-to-text if captions aren’t there), run the text through an AI trained to spot these reasoning tricks, and overlay the results on the video player in real time. Start simple with YouTube and the most common fallacies, then grow it into a tool for all major video platforms.”


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

Experienced contributors, what is something you would tell yourself to do sooner if you were starting out again?

Title, looking for learnings or suggestions on the open source journey

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

Meta Platform

Hello Guys!

I’ve wondered why everybody in my area uses Meta’s Messenger and Facebook instead of Signal, Element, or others. They are much more private, lighter, and easier to use. I couldn’t persuade anybody. If you had to convince your family members, how would you do it?

Thank you for your reply.

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

A new experiment: making Protobuf in C++ less painful (inspired by the old “why is Protobuf so clunky?” thread)
https://github.com/illegal-instruction-co/sugar-proto

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