UnisonDB - Log-based real-time database

Log-native real-time database : Designed to solve data consistency and real-time responsiveness challenges

* Combining a B+Tree storage engine and Write-Ahead Logging (WAL) -based streaming replication, it ensures sub-second replication and strong consistency across hundreds of nodes.
* Supports Key-Value, Wide-Column, and Large Object (LOB) storage with a multi-model structure
* Optimized for local-first architecture with edge-first design
* Multi-tenancy support through namespace isolation

Differences from existing systems:

* LMDB/BoltDB is a fast local storage but cannot be replicated.
* etcd/Consul has high consistency but limited scalability.
* Kafka/NATS is strong in streaming, but not queryable.

UnisonDB bridges this gap, providing a single, log-centric architecture that integrates storage and streaming.

Core architecture 3-tier structure

1. WALFS (Write-Ahead Log File System) * – mmap-based log file system, optimized for large-scale read/write
* Segment-based log structure, optimized for both sequential writes and random reads
* Supports zero-copy reading , offset-based seeking , and real-time tailing
* Parallel reader architecture that allows multiple replica nodes to read simultaneously

2. Engine – Hybrid storage combining WAL, MemTable, and B-Tree
* Combination of MemTable (skip list) and B-Tree index based on WALFS
* Transfer without deserialization when replicating using FlatBuffers
* Supports atomic multi-key transactions , ensuring consistency at the commit level.
* LOB (Large Object) can be chunked and streamed in units of transactions.
* Wide-Column model supports partial column updates and dynamic schema expansion.

3. Replication – WAL-based streaming replication, including offset tracking.
* WAL-based streaming replication , where followers track offsets and synchronize in real time.
* Maintain self-describing data structures using FlatBuffer log records .
* Efficient streaming implementation with batch transmission
* Consistency-focused design

UnisonDB's solution
* Append-only log + B-Tree combination provides high-speed writes and efficient range reads.
* Support for transaction-based multi-key replication and column-aware synchronization
* Built-in replication with gRPC WAL streaming + B-Tree snapshots

License : Apache License 2.0

Development language : Go
https://github.com/ankur-anand/unisondb

https://redd.it/1p404fi
@r_opensource

I built the tiniest task manager
https://github.com/denisbondare/taskbucket

https://redd.it/1p42l8g
@r_opensource

Open sourced my coding problem typing trainer. Looking for contributors or feedback on code structure
https://github.com/cwklurks/codesprint

https://redd.it/1p42ajq
@r_opensource

Seeking feedback on my experimental js lib oem

I've been building and rebuilding a framework off and on for a couple years. I recently had an ah-hah moment and reworked things to a 2.0 version. I just posted the new version here: https://oem.js.org/. I'm curious what people think. The core idea is that it's a framework to design your own framework. It's only 300 LOC and it facilitates a particular syntax for your own framework that results in code you can understand from top to bottom.

https://redd.it/1p41o7s
@r_opensource

Somachord: A Subsonic client (Web, Electron, PWA)

https://github.com/sammy-ette/Somachord

Somachord is a web music player for Subsonic servers (and subsonic servers only) written in Gleam :)

It works on both desktop in the browser and with Electron, and mobile with a PWA as well

https://redd.it/1p48bu6
@r_opensource

Writingway 2: Scrivener meets AI-integration

I wrote a freeware version of sites like NovelCrafter or Sudowrite. Runs on your machine, costs zero, nothing gets saved on some obscure server, and you could even run it with a local model completely without internet access.

Of course FOSS.

Here's my blog post about it:
https://aomukai.com/2025/11/23/writingway-2-now-plug-and-play/

https://redd.it/1p4atb9
@r_opensource

ORION: A free, open-source tool to scrape and map real-world engineering career paths
https://mrweeb0.github.io/ORION-tool-showcase/

https://redd.it/1p49qhz
@r_opensource

Secure storage for chat data & its workloads
https://github.com/progressdb/ProgressDB

https://redd.it/1p4e3jh
@r_opensource

I rewrote the UI in Vue.js for Go benchmark visualization
https://github.com/goptics/vizb

https://redd.it/1p4g6xw
@r_opensource

I built a free scanner to check if your website is i18n-ready

I realized most websites have broken or missing internationalization setups, no lang attribute, wrong hreflang, untranslated strings, etc.
So I built a free scanner that analyzes any website and gives an i18n readiness score with a few SEO insights.
It’s a small tool I made to help devs see if their site is ready for global users.

👉 Try it: https://intlayer.org/fr/i18n-seo-scanner

Feedback welcome especially on the checks or UI!

https://redd.it/1p4hk91
@r_opensource

Need honest feedback on my AI workflow library (2 months of work, feeling stuck)

Hey everyone,

I've spent the past 2 months building a TypeScript library for creating AI workflows that process data in steps. The core idea:

Each step can process a section of data or work globally (waiting for all sections of data to be done on the global step)
Steps can depend on other steps
You define prompts, dependencies, and data transformations, steps ai model configurations
16 Hooks let you inject async integrations at any point in the workflow

My problem: I finished it, but I'm not happy with the result. It still requires too much boilerplate and infrastructure code. My original vision was something where you just configure prompts and dependencies - minimal code, maximum clarity.

I'm too close to this project now and don't have a realistic view anymore.

What I'm looking for:

Honest critique of the concept itself
Is this even solving a real problem?
If you check out the repo, what would you change?
Interested in collaborating to make this actually useful for the TS/AI community?

GitHub https://github.com/dagengine/dagengine
Homepage: https://www.dagengine.ai/ (You can check examples and documentation)

I'm genuinely open to pivoting the whole approach or scrapping it if it's not the right direction. Just want to build something people will actually use.

https://redd.it/1p4k3ur
@r_opensource

go-agent

Take a look at go-agent.

It's new agentic framework in golang that supports utcp and codemode, agents as tools, has its own memory layer. It supports many llm providers.



Repository:

https://github.com/protocol-lattice/go-agent

https://redd.it/1p4k0be
@r_opensource

Opinions on this Project

Hey, first time on this subreddit, I marked this post as promotional simply because I’m not sure what the rules are here, but really I’m just looking for helpful, constructive feedback on a Java plugin library I built. Please let me know what you think ;p

GitHub: https://github.com/sieadev/Jonion
Docs: https://docs.siea.dev/jonion/

https://redd.it/1p4ja2n
@r_opensource

Open source roku alternative

This question has been asked a couple times but it's been a few years and there were never any good answers from what I've seen. I'd like to replace my roku with something with better privacy and no ads. I don't have a media library that I'm trying to host and I'm not looking to pirate, I just want something I can connect my streaming services to. It seems like it shouldn't be that difficult but I haven't found a good solution. My thought is that android can install and run apps like youtube, netflix, etc, so shouldn't there be a way to install an android distro on a raspberry pi or something and download apps onto it? Not sure if there's a project like this or if someone has looked into it. Thanks.

https://redd.it/1p4q25h
@r_opensource

I built this free yoga app, open to contributions!
https://simonschubert.github.io/YogaBase/

https://redd.it/1p4qjfj
@r_opensource

Building an open source AI powered DB monitoring tool

Hi folks,

I am building an open source platform to monitor all your SQL, NoSQL and KeyVal DB at a single place and it gives you insights on how you can optimize your DB.

Do you think any other OSS tool does that?

Code: github.com/meshag-ai/monitor

https://redd.it/1p4uj58
@r_opensource

Built a webapp that transfers your Spotify playlists to YouTube Music: no limits, login optional
https://github.com/OwaisSafa/melody-shift

https://redd.it/1p4wkyz
@r_opensource

wayscriber - annotation & whiteboard app for Linux
https://github.com/devmobasa/wayscriber

https://redd.it/1p4xpw7
@r_opensource

I build GitViews: a simple view counter for your GitHub profile and repos

Hello everyone!

I recently came across the GitHub Profile Views Counter project on GitHub and decided to try it, only to find out that it doesn't support separate counters for individual repositories, something I needed. As the name suggests, it's intended only for the profile README.

So, I decided to build one for myself that supports both a profile counter and per-repo counters, as well as a badge showing the sum of all repository views. It's called GitViews. It's open-source and free to use.

I’d love to hear your thoughts and any suggestions on how I can improve it!

https://redd.it/1p4wf0m
@r_opensource

I made a one-click macOS batch video compressor

Hey folks,



I’ve been wrestling with messy video folders for a while, so last weekend I finally sat down and built a small tool to make that a bit less painful.



It turned into this:



HandBrake Batch Compressor (HBC)

👉 https://github.com/kemalsanli/HBC



It’s a tiny macOS app (SwiftUI) that tries to make batch compression as simple as:



>pick a folder → click once → let it chew through everything



What it does:



Recursively scans a source folder (and all subfolders) for video files
Uses a command-line encoder to compress them in batch
Has a safe mode:
Writes all compressed files into a separate compressed folder that mirrors the original structure
Leaves your originals exactly as they are
Has an optional YOLO mode for people who prefer more automation:
For each file, if the new one is smaller, it replaces the original
If it’s not smaller or encoding fails, it keeps the original
So it still has some built-in safety, it’s not a blind “delete everything” switch
Includes an “Optimize Original Folder” pass for the more cautious / control-freak workflow:
You first run in safe mode and let it build a compressed folder
You can review the results there
When you’re happy, HBC can walk through that compressed folder and only replace originals when the compressed version is actually smaller
So it becomes a two-step process: first generate results, then selectively apply them back into your archive
Can write a run log into the source folder so you can see exactly what happened (size comparisons, replacements, errors…)





A few notes:



It’s fully open source and completely free – no Pro version, no paywall, no tracking
It’s not tied to any commercial product; it just leans on a CLI encoder for all the heavy lifting under the hood





I mostly built this for my own archive, but since it’s open source:



If you want to review the code,
suggest cleaner patterns / better defaults,
or add translations / improvements,





I’d really appreciate any feedback, nitpicks, or PRs.

https://redd.it/1p4zoti
@r_opensource

An open-source CLI tool with a TUI dashboard for monitoring services

https://github.com/abhixdd/UptimeKit-CLI

https://redd.it/1p4vpav
@r_opensource