I worked on an online compiler , support 20+ languages


Hey everyone!

I'm building Codeer.org, a fast and simple online code compiler where you can write and run code instantly in your browser.

I'm looking for people who'd like to collaborate whether it's adding features, improving the Ul, or testing. All skill levels welcome!

I KNOW SOME MAY NOT WORK PROPERLY, JUST TRY GUEST MODE 😅

Demo: https://www.codeer.org/

GitHub: https://github.com/siddu-k/codeeride

https://redd.it/1muir6h
@r_opensource

Full open-source dashboard template
https://template.openstatus.dev

https://redd.it/1mukzy3
@r_opensource

Open sourced a CLI that turns PDFs and docs into fine tuning datasets now with multi file support

Repo: https://github.com/Datalore-ai/datalore-localgen-cli

Hi everyone,

During my internship I built a small terminal tool that could generate fine tuning datasets from real world data using deep research. I later open sourced it and recently built a version that works fully offline on local files.

I shared this update a few days ago and it was really cool to see the response. It got around 50 stars and so many thoughtful suggestions. Really grateful to everyone who checked it out.

One suggestion that came up a lot was if it can handle multiple files at once. So I integrated that. Now you can just point it at a directory path and it will process everything inside extract text find relevant parts with semantic search apply your schema or instructions and output a clean dataset.

Another common request was around privacy like supporting local LLMs such as Ollama instead of relying only on external APIs. That is definitely something we want to explore next.

We are two students juggling college with this side project so sorry for the slow updates but every piece of feedback has been super motivating. Since it is open source contributions are very welcome and if anyone wants to jump in we would be really really grateful.

Thanks again for all the support and keep the suggestions coming. It has been amazing building this with input from the community.

https://redd.it/1muil3e
@r_opensource

Wikipedia graph explorer

hi there

i made a Wikipedia graph explorer https://wikigrapher.com/

its still in early stage with more enhancements to come

graph generator is shell/python

back/front is spring webflux + vanilla html/TS

i would be more than happy if anyone is interested in giving feedback or collaborating



https://redd.it/1muobp8
@r_opensource

New here, would like if y’all can give suggestions or maybe contribute

Hi so i’m deep invested in ai assistants and agentic ai. What i’ve got so far is a floating chat bar with ollama as backend. My vision is to make it agentic in the long run where user can call it using global hotkeys and give tasks. I’m more inclined towards macos agent.

https://github.com/abhaymundhara/ChatDock

Would like some honest feedback or responses.
Thank you for your time.

https://redd.it/1mup8ic
@r_opensource

What are you guys building for the open source community?
https://trysita.com

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

Open-Source Docs Platform for IT Pros w/ AI Search & Collaboration

I’m building Surge, an open-source documentation platform for IT teams, offering enterprise-grade features without vendor lock-in. It combines AI-powered search, robust collaboration, and strong security to simplify knowledge management.

Key Features:
🚀 Collaboration: Real-time editing, threaded comments, version control with diffs.
🏢 Workspaces: Multi-workspace setup, role-based permissions, analytics, custom branding.
🤖 AI Search: Natural language queries, smart suggestions, OCR processing.
🔒 Security: SSO/LDAP, granular access, audit logs, secure doc vaults.
📱 UI: Mobile-optimized PWA, offline support, rich-text editor with diagrams.

Links:
GitHub: https://github.com/SteffMet/Surge
Demo: https://demo.bitsurge.io


Early stage project—would love your thoughts on it.

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

Youtube on Smart TV? Freetube (PC), NewPipe (Android) integration

On PC, I use FreeTube client, on Android I use NewPipe/Tabular. I have a smart TV with an Android TV box.

I would like an easy way to send Youtube video links/playlists from Android phone to play on the smart TV, is this possible without using a third-party service? I would much prefer not to mirror phone screen to TV for this purpose (potentially drains battery life, unwanted notifications showing on screen, etc.).

Is there a way integrate between all 3 clients (Freetube, NewPipe, and whatever client to use for the Android TV box)? Ideally, they share the same subnoscriptions, playlists, etc. The clients don't seem to have a way to be able to sync these together nor do they support a locally hosted account that is comppatible with each other.

I use Syncthing which can sync data between devices, but since they don't use the same formats, it's not of much help. Freetube/Newpipe and presumably SmarTubeNext for the TV are great apps on their own, the is there really no integration between them such that you seamlessly switch between them with auto-synced subnoscriptions and playlists?

My workaround is to download the videos on a Jellyfin server and play from there but obviously it requires storage space and seems too expensive for what is usually "come across a video of any interest, download, maybe watch once, delete".

I'm not sure if KDE Connect lets me copy Youtube link to clipboard and share it to Android TV box, but it'd be annoying because it also requires manually pasting to Youtube client to play. It would be nice of Jellyfin allows presenting Youtube links/playlists as if it's part of the library.

https://redd.it/1muv7xd
@r_opensource

Qwen-Image-Edit available
https://x.com/Alibaba_Qwen/status/1957500569029079083

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

Syncing with Postgres: Logical Replication vs. ETL
https://www.paradedb.com/blog/etl_vs_logical_replication

https://redd.it/1mut73q
@r_opensource

Built an open-source-inspired tool for Fediverse posting: loftly.social

I couldn’t find a good way to post across Mastodon, Bluesky, Pixelfed, and Lemmy simultaneously, so I built one: Loftly.social
It’s live now, and I’d love to hear from the open-source community—what features or improvements would make this most useful to you?

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

How to actually understand large code bases and Start working start

I am new to open-source and I am trying to understand large codebases and it's really difficult for me. I am getting nowhere just looking at it.

https://redd.it/1mv3sqk
@r_opensource

DeepSeek v3.1 just went live on HuggingFace
https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Base

https://redd.it/1mv4prs
@r_opensource

GhostDeck Update (Multipurpose Utility Deck)
https://youtu.be/9lz6rsmdUvs?si=ENWJ1H1WKS0L0Ftb

https://redd.it/1mv4dwl
@r_opensource

Stop Paywalling Security: SSO Is a Basic Right, Not an Enterprise Perk
https://oneuptime.com/blog/post/2025-08-19-sso-is-a-security-basic-not-an-enterprise-perk/view

https://redd.it/1mv8msa
@r_opensource

BlinkDB - Time Traveling, in-memory KV database

Hey folks,

I’ve been messing around with Go lately and ended up building **BlinkDB**, a little in-memory key-value store. Think of it as Redis’s younger cousin who’s obsessed with time travel.

Right now it does the basics: `Set`, `Get`, `Delete`. On top of that, it supports TTLs (both relative and absolute), a compare-and-set for safe updates, and an append-only history log so you can ask *“what did this key look like yesterday at 3pm?”* with `get_when`.

There’s also a `SweepExpired` to clean up dead keys — though I’m leaning toward making them tombstones instead, so the history stays consistent. Aside from adding mutexes for concurrency, it’s basically done for v1.

A couple things I think are fun:

* You can literally rewind state like a DVR.
* Everything is just stdlib Go — no giant deps.
* It’s simple to hack on, but could evolve into persistence, clustering, or even a toy Redis-style CLI.

Repo’s here if you want to poke around:
[BlinkDB](https://github.com/ARJ2211/BlinkDB)

And if you think it’s neat, tossing a star my way would be awesome :)

https://redd.it/1mv94x1
@r_opensource

Omaro - TUI for browsing posts and comments on lobste.rs
https://github.com/Rolv-Apneseth/omaro

https://redd.it/1mva564
@r_opensource

Need your opinion on my open-source language syntax

Hey everyone,

I’m currently working on an exciting new project called DeduKt, a symbolic computation language designed for use in scientific research and complex computations. One of the core aspects of developing DeduKt is ensuring the syntax is as intuitive and effective as possible for users in the scientific community.

To make sure DeduKt is a language that truly serves its purpose, I’ve put together a survey focused on gathering your opinions about the preferred syntax for scientific computing. It only takes a few minutes, and your feedback would be incredibly valuable.

If you’re interested in contributing, please take the survey here: https://form.typeform.com/to/g8yi9oTn

Thank you so much for your time and support in shaping the future of DeduKt!

Cheers,

https://redd.it/1mvb5ku
@r_opensource

LunarBase - Security first portable BaaS

I have created a new project - LunarBase. It is hosted on its own server, a single binary, providing a database management platform where security is put first. Each component is designed to protect data while maintaining real-time capabilities.

Key features include:

• ⁠password hashing using Argon2id,
• ⁠dynamic JWT,
• ⁠multi-level access control,
• ⁠database encryption using SQLCipher,
• ⁠real-time subnoscription system using WebSocket.

The frontend is based on my own Nocta UI component system in a copy-paste philosophy with full TypeScript support. The technology stack is Rust + Axum + Diesel on the backend and React 19 + TanStack Router on the frontend. The whole thing compiles to one single binary with embedded resources, which greatly simplifies deployment. I'm most proud of the granular permissions and overall security approach.

This was a big lesson in practical Rust for me. The code is open source, so you can see how I approached various problems. I encourage you to contribute - any help, whether in reporting bugs, adding new features or improving the documentation, is welcome.

On the project's roadmap in the near future are modifying the permissions system to make it more intuitive, building in a rich text editor for an easy way to write blog posts and embedding JS engine (probably BOA) to allow users writing their own JS API extensions

Repo URL: https://github.com/66HEX/lunarbase

https://redd.it/1mvd1dn
@r_opensource

New TilBuci version, a free software for interactive contente creation

TilBuci, a free software (MPL-2.0 license) for the creation of interactive digital content for the web, apps and the like, reaches version 14. To check it out, access the software repository at

https://github.com/lucasjunqueira-var/tilbuci/releases/tag/v14

New features

Text files

Support for a new type of media file has been added, “string media files”. These are files in JSON format that can be loaded and unloaded at any time into variables, allowing your creations to contain large volumes of text with reduced impact on load time and memory usage.

Workspaces

Until now it was only possible to edit one movie/scene at a time. The new "+Workspace" button significantly improves the usability of the software, allowing you to edit multiple scenes and even multiple movies simultaneously in single or multi-user installations.

Portable desktop versions

TilBuci is a web software with several multi-user features for collective creation. However, there are cases where local use by just one person may be necessary. With that in mind, we now have a desktop version, presented as a portable software that can be copied to your computer or even to external drives, without the need for installation. The portable version is available for Windows, Linux, and macOS (x64-based architectures). Note that when performing in this way, TilBuci's server functions, such as visitor identification (login) or cloud data storage, will not be available, but creations made in the desktop version can be easily exported and imported to a server installation in the usual way.

Next steps

For the next versions, features are being worked on to simplify the creation of narrative content, such as "visual novels". The planned tools include character registration, dialogue generation and display (inspired by the Renpy engine) and definition of multilinear narrative structure (inspired by the Twine tool). In addition, an exporter for "activities" on Discord is in development.

About TilBuci

TilBuci is an interactive content creation tool focused on development for web, mobile and desktop apps. Distributed as free software under the MPL-2.0 license, it is presented in the form of a web program, executed from a browser with functionalities for collective creation, and also as a portable desktop software for various systems. To learn more about the project, visit tilbuci.com.br .

https://redd.it/1mvdnv5
@r_opensource

IBM and NASA just dropped Surya — an open‑source AI to forecast solar storms before they hit

Solar storms don’t just make pretty auroras—they can scramble GPS, disrupt flights, degrade satellite comms, and stress power grids. To get ahead of that, IBM and NASA have open‑sourced Surya on Hugging Face: a foundation model trained on years of Solar Dynamics Observatory (SDO) data to make space‑weather forecasting more accurate and accessible.

What Surya is

A mid‑size foundation model for heliophysics that learns general “features of the Sun” from large SDO image archives.

Built to support zero/few‑shot tasks like flare probability, CME risk, and geomagnetic indices (e.g., Kp/Dst) with fine‑tuning.

Released with open weights and recipes so labs, universities, and startups can adapt it without massive compute.

Why this matters

Early, reliable alerts help airlines reroute, satellite operators safe‑mode hardware, and grid operators harden the network before a hit.

Open sourcing lowers the barrier for regional forecasters and fosters reproducible science (shared baselines, comparable benchmarks).

We’re in an active solar cycle—better lead times now can prevent expensive outages and service disruptions.

How to try it (technical)

Pull the model from Hugging Face and fine‑tune on your target label: flare class prediction, Kp nowcasting, or satellite anomaly detection.

Start with SDO preprocessing pipelines; add lightweight adapters/LoRA for event‑specific fine‑tuning to keep compute modest.

Evaluate on public benchmarks (Kp/Dst) and report lead time vs. skill scores; stress test on extreme events.

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