alt Hacker NewsAsk HN: What Are You Working On? (April 2026)
What are you working on? Any new ideas that you're thinking about?
Comments
Handcrafting a new version of Collider.JAM (http://collider.land/) - a JavaScript framework for rapid game prototyping. Ludum Dare 59 is coming, and any coding samurai needs to sharpen their katana. And Ludum Dare is a perfect excuse to do just that.
IMHO, Ludum Dare is The Jam for many reasons and it is a pity that it's slowly fading away (the last Mike Kasprzak's post explains some of the challenges https://ldjam.com/events/ludum-dare/59/$424396/ludum-dare-59...). It looks like we don't have many events left in the pipeline, and this one will be a nice opportunity to participate and still enjoy the sunset of the era and the vibes of the awesome community.
I am mostly polishing old code and old documentation, nothing fascinating. For instance yesterday I fixed a small project called "environment", or fixed it to about 90% - output example here:
https://i.imgur.com/uHpnUox.png
(Don't mind the errors or display issues for now; this project will report which versions of programs are installed on the given computer, and which ones could be updated. A few smaller bugs still remain but the main rewrite was finished now.)
For work-related reasons I have to expand on some python tasks in the coming weeks. Trying to find something interesting here, but I guess I'll just focus on various bioinformatics-related tasks (also related to work). Programming for the most part is not extremely interesting - the only part I actually like is any creativity and useful end results. Fixing bugs is annoying to no ends. Writing documentation is also boring, but can not be avoided.
I'm building Repple (https://repple.sh), a free flashcard app with spaced repetition scheduling and some (tasteful imo) AI additions. I've been using Anki for 10+ years and love it but always wanted something with a cleaner UI and that's a bit easier to experiment with. The recent Anki ownership change pushed me to finally make something.
Added a REST API (https://repple.sh/developers) a few weeks ago so you can build on top of it. Decks, cards, reviews, etc.
Feature delta over Anki:
- Tab-autocomplete for text fields
- Automatic image-gen for image fields
- Optional rephrasing that changes wording each review to avoid pattern matching
- Basic PDF library & incremental reading support
- "Orphan" card detection; i.e. knowledge that isn't well connected
- ... + a bunch of other qol improvements like semantic search, etc.
LLM has made scripts incredibly cheap, and their lifecycles as short as one-off. Batch rename? "Please implement a Python script." Remove background from images? "Please implement a Python script." Or various operations that could be described in a few sentences but used to take a lot of time—"help me implement a script..." With development time nearly zero, creating a new file, running a script, then deleting the script becomes the most time-consuming part, which feels very clunky. So I wrote RunOnce—targeted at this kind of one-off script scenario. It registers in Windows 11's right-click menu; click "Run Code Here" and a minimal editor appears. Paste your code (or generate it inside), run—automatic language detection, file cleanup, etc., much smoother :) Written in WinUI3, follows Windows 11 Mica guidelines, distributed on the Microsoft Store: https://github.com/Water-Run/RunOnce
CnakeCharmer - https://github.com/dleemiller/CnakeCharmer
https://huggingface.co/datasets/CnakeCharmer/CnakeCharmer
This project started from a belief that llms should be better at doing python to cython code translations than they are. So we started setting a large set of parallel implementations.
Then I realized that Claude code was much better at working on the data using tools (mcp) to check and iterate. The scope transformed into an platform for creating the SFT agentic trace dataset using sandboxed tools for compilation, testing, linting, address sanitizing and benchmarking.
We still need to bulk up the GRPO dataset with a large number of good unmatched python examples. But early results using SFT only on gpt-oss 20b are quite good.
I've never been a pro gamer but I've spent an embarrassing amount of time min-maxing graphics settings — chasing that sweet spot between FPS and visual fidelity. I built FragForge (https://fragforge.app/) to scratch my own itch: one place to find optimal per-game settings tuned to specific hardware.
The key insight I found was that the best real-world knowledge base for this already exists — Twitch streamers. Many show their settings on stream, and they're running known hardware. By extracting settings from clips and pairing them with each streamer's setup I've built a database of 200+ curated configs that go way beyond "just set everything to low" and looking to add more.
If you've ever burned 30 minutes tweaking shadow quality and anti-aliasing before actually playing a game, this is for you. Would love any feedback.
I've been working on proving that Claude Opus can be self-reflecting meaning that its attention and context are large enough that it is aware of its own instructions, aware of the task, and capable of writing its own instructions to optimize solving the task in recursive iterations. [0]
By tuning the agent, it is possible to create trading strategies [1] and reverse engineer websites in order to create optimized JSON APIs using the websites internal private APIs. [2]
I'm having the hardest time communicating what is happening so next I'm going to try to explain it using data visualizations so people can visualize it in action.
[0] https://github.com/adam-s/agent-tuning
[1] https://github.com/adam-s/alphadidactic
[2] https://github.com/adam-s/intercept?tab=readme-ov-file#how-i...
I’m building https://connections-hint.today, a spoiler-safe helper for NYT Connections.
The idea is simple: hints first, answers later. It has daily updates, archive pages, Sports Edition support, and a lightweight analyzer for reviewing misses.
Built with Astro + Cloudflare Pages/Workers + D1.
I am working on making grocery online shopping less overwhelming and more like a rolling list, you keep adding items as you see them (missing) in your household and it silently records it at the backend. When you are ready to pick up order, you push to qfc cart via api (a button) and boom your grocery shopping is done. No need of making lists and then one by one putting them on the cart. It works with any QFC or Kroger store because to my disbelieve they actually have an open sku and cart api. Grateful to Kroger to be tech forward. Free to use , here is the link https://www.ddisco.com/sonic/customer My wife is hooked on it as she had to take time in the week to sit down ask me what to order and then build the cart. Now it’s like just typing in what you need.
Next I am making the version for folks who do not make a list and just go with past orders , for them I am automating so the cart is made based on past orders like milk usually is ordered every 2 weeks.
Building Lexplain (https://lexplain.net/release-notes), an AI-powered service that explains Linux kernel commits and release notes.
Most engineers running Linux in production aren't kernel developers. Keeping up with kernel changes is hard, and unexpected kernel behavior silently impacting production systems happens more often than it should.
So I wanted a way for those engineers to scan through kernel changes quickly. Not just what lines changed, but what the code actually does, why it matters, and what it means for real systems. Something closer to what art museum docents do.
7,000+ commit analyses and release notes for the 7.0-rc series are available. Release notes for 7.0 stable are in progress.
While playing around with OpenClaw I realized that after a few days of adding skills, plugins, crons, etc. things were all over the places. Memory files spread everywhere in the workspace, plugin configs stored here and there, skills not always visible to crons, ...
I started working on a CLI to keep OpenClaw organized: https://clawtique.ai
The analogy is that of a boutique. OpenClaw goes to the boutique and is "dressed up" properly so that all the various components are organized and easy to maintain. Clawtique is organized around the concept of a "dress", basically a bundle of everything OpenClaw needs to achieve a goal (skills, plugins, memory segments, crons, ...). The CLI enables users to easily dress and undress OpenClaw so that you can try out a dress and easily remove it without leaving any dangling dependencies.
Some dresses i created are the sleeping-coach (bundles the OuraClaw plugin https://github.com/rickybloomfield/OuraClaw with skills and crons that notify you on how you slept) and the fabric-sync (bundles the fabric plugin https://github.com/onfabric/openclaw-fabric with skills and crons to maintain an accurate USER.md of you based on your interactions on the various big tech platforms).
The follow up is to have OpenClaw use the Clawtique CLI itself so that it can easily dress and undress with whatever it needs to accomplish the goal without everything becoming an unmanageable mess.
Here is the repo: https://github.com/onfabric/clawtique
Curious to know what you guys think
Building MyLawnGuide (https://mylawnguide.com) as a way to help me organize and plan my yard maintenance for the season.
Enter your zip, type of grass you want, where you're starting and it gives you a rough plan to follow for seeding, feeding, and pre-emergent.
The workflow orchestration tools I've used over the past have been consistently unsatisfying, so I've been building my own. With Coflux (https://coflux.com), workflows are defined declaratively in plain Python with decorated functions (tasks). Workers connect to the orchestration server and get assigned work - calls to other tasks get intercepted and then re-scheduled by the server onto (potentially different) workers. Tasks are executed in pre-warmed, isolated processes, with low latency.
Beyond standard features (retries, caching, timeouts - enabled with attributes on the decorator), Coflux supports more novel features - like suspense (where a task can choose to go to sleep and get restarted when a result it depends on becomes available), memoisation (where steps within a run are aggressively cached so that you can re-run steps in a workflow without re-running upstream steps), and the ability to re-run a step in a different workspaces (with updated code, or in a different environment).
It turns out this works great for implementing agentic systems - you can provide references to tasks as tools to an LLM call and have the AI drive - tasks can be easily sandboxed. And Claude is very capable of using the CLI to interact with the orchestration server to submit workflows, investigate failed runs, make updates to workflows and re-run steps.
I'm trying to make sure it's easy to try out - there's a self-contained CLI that can be used to start the server (a single Docker container), run worker processes, and then interact with the server. The dev mode automatically restarts the workers as you make local changes. There's also a hosted UI for observing runs in real-time, where you can see the execution graph, access logs/metrics/assets/etc - it works without creating an account - the browser interacts with your orchestration server directly.
I'm building a hiring platform for hardware engineers, contrary to hiring software engineers where there are too many options, hiring hardware engineers through provisioned testing environments lacked some good software, a good remote-assessment tool.
A MVP is live at this https://ai-eval-lab.janardan.xyz/
Github: https://github.com/janardannn/ai-eval-lab
For now, for every session, inside the docker container, i create a virtual frame buffer using xvfb and render KiCad in it. Thereafter using x11vnc to pickup the xvfb display and serve it using VNC. Since browsers cant use raw TCP connections, i am using websockify to bridge WS <--> TCP. To render the app in browser, I am using noVNC to render an iframe inside my app. noVNC opens a websocket to websockify and the frames from the container start flowing.
Inside, KiCad itself, a have added a python plugin which auto loads when KiCad is started as a background thread. After a set interval of time, it tracks all the board objects like footprints, tracks, zones etc and serialize them as a JSON and makes a POST request to the backend server for the specific session id.
For the grading and evaluation, as of now, I am using an LLM to cross check the entire sequential board activity that was sent over the duration of the assessment and grade it against a predefined rubric/score it against per-checkpoint criteria. This is entirely based on how the admin have set the assessment grading rubrics.
I am not building it as a business rather exploring the space and the technology. It was an excuse for me to learn about VNC, streaming, docker orchestration and also somewhat of AWS through its deployment.
I'm hosting it on the free AWS tier so only one session at a time. If you queued then please give it some time.
Happy to chat about anything regarding the project!
Hey all, I’ve built Nekode (https://github.com/Jakob-98/nekode), which is desktop-cats for monitoring your AI coding agent sessions (OpenCode/Claude Code/ Copilot CLI/VSCode Copilot). It works as a macOS menubar app, and the cats run around on your desktop.
Each cat mirrors the agent's state, such as sleeping when idle, walking when working, sitting when waiting for input, running toward your cursor when it needs permission.
Fully native Swift, no Electron, under 5 MB, zero network requests, all session data stays local as plain JSON.
I published it source-available with an honor-system license, but this week I’m going to fully open source it and remove the licence. The payment/nag system was an interesting experiment but the project is more useful to me as a proper OSS tool at this point.
I'm building Styleframe (https://styleframe.dev), a zero-runtime CSS-in-TS Framework written for theming and maintaining Design Systems.
After 4 years of maintaining an Open Source Design System, I needed a better way for theming than Sass and PostCSS. I needed the power of a full-featured programming language. That's how the first version of Styleframe was born.
My vision is for Design Systems to be endlessly configurable and composable, like you would configure any library, with or without AI. Want to change your entire website to look like Linear? Simply install and use the Linear Design System configuration. Want only your buttons and cards to look like Linear, and the rest be the default theme? Use the button and card composable functions from that package.
Styleframe is built as a transpiler-first system. You write your design tokens, selectors, utilities, and recipes in TypeScript, and Styleframe tokenizes everything into an internal representation. From there, the transpiler generates dual outputs:
- CSS output: variables, selectors, utilities, themes, keyframes
- TypeScript output: typed recipe functions with full IDE autocomplete (with an optional Runtime)
This architecture means you can have complete control over customizing how your system is output. You could even use the generated tokens to render documentation for your design system components. The output code can be integrated with any headless UI Components Library, or with your own custom components.
Fun fact: I've reimplemented the entirety of TailwindCSS using Styleframe's utility and modifier tokens. Not only that, but I've also built a scanner which picks up your CSS Classes from your markup. It's basically Tailwind, but 100% configurable (even utility class format can be changed), and is always based on your design tokens.
I'm building Envelope, which is a checking account with built-in budgeting: https://envelopebudgeting.com It's like if Mint, YNAB, or Copilot was built into a bank account. The goal is to help people budget, spend, and save from one account, instead of using a separate budgeting tool, spreadsheet (yuck!), or cash envelope system. It does that by allowing users to get paid via direct deposit and then organizing their money into digital envelopes for bills, groceries, savings, sinking funds, etc. It's attached to a debit card so everything updates instantly and has individual or joint accounts.
I'm currently working on the paycheck prediction algorithm so Envelope can automatically determine how much needs to be set aside from user's paychecks so their bills are paid on time.
Last month I released my first Steam game, an occasionally frustrating game about scoring great goals: https://store.steampowered.com/app/3802120/Put_One_In_for_Jo....
I built HeartRoutine to help me lower my LDL and ApoB. I recently started beta testing it on some friends, too, to see if anyone other than myself would find it helpful: https://www.heartroutine.com/.
I've started building the combat prototype for my next game, Today I Will Destroy You, inspired by my love of going-on-an-adventure-with-a-sword games and Sekiro-style combat.
I've committed to keeping my personal website up to date: https://piinecone.com/.
I am working on a system built around the OpenAI Responses API WebSocket mode as performance is something that interests me.
Its like a microservices architecture with NATS JetStream coordinating stuff. I want to keep the worker core as clean as possible, just managing open sockets, threads and continuation.
Document querying is something I am interested in also. This system allows me to pin a document to a socket as a subagent, which is then called upon.
I have hit alot of slip ups along the way, such as infinite loops trying to call OpenAI API, etc ...
Example usage: 10 documents on warm sockets on GPT 5.4 nano. Then the main thread can call out to those other sockets to query the documents in parallel. It allows alot of possibilities : cheaper models for cheaper tasks, input caching and lower latency.
There is also a frontend
Alot of information is in here, just thoughts, designs etc: https://github.com/SamSam12121212/ExplorerPRO/tree/main/docs
Just today I helped co-host the SF Permacomputing Club.
It was a lot of fun and I love all the good energy people bring to the conversation about long lasting and community driven tech.
Looking for an alternative to HN, I decided to spend 6 months writing some code and got my first result, which is how https://comuniq.xyz was born. It's nothing compared to the power of Hacker News, but it could become an alternative. To make it stand out, I added notifications, Markdown text styling, and a layout that the community there says is pretty nice.
Just recently I finished a AI agent skill for making wireframes and clickable prototypes in the hand-drawn aesthetics (Github: https://github.com/agilek/wireframer-skill)
Sample output:
- https://agilek.github.io/wireframer-skill/samples/dashboard....
- https://agilek.github.io/wireframer-skill/samples/bio.html
- https://agilek.github.io/wireframer-skill/samples/checkout.h...
Playing with an idea of a next-gen self-hosted media server software, with rust, svelte and all the goodies.
But at my current knowledge and practical work, its like giving a chimpanzee a nuclear reactor schematic. But it's a passion project idea of mine, I really want it to become real one day. Personally, I feel like something much better can be made than current solutions.
Im working on a all-in-one event management platform for studios, event planner etc..,
It consists of CRM, Expense tracking, Equipment Management, Event Gallery( photo share, Face Detection based download, Guest Upload) etc..,
Currently working on moving it from cloud supabase to self hosted version.
Creating interactive pivot tables from large relational tables.
Many people know that a handy data analysis feature in Excel is to create a pivot table from a spreadsheet. But spreadsheets are limited to just a million rows. You can get around this limit by jumping through a bunch of hoops.
My system lets you easily create tables with thousands of columns and hundreds of millions of rows. (Just drop a CSV, Json, or other file on a window to create a table.)
Now you can create a pivot table from it with just a few clicks of the mouse. It is fast (I created a pivot table against an 8.5 million row table of Chicago crime data in less than a second.)
The resulting pivot table is interactive. Each cell (row/column intersection) has all the row keys mapped to it. Double-click on any cell and it will instantly show you all the rows in the original table that were used to calculate the cell. You can then analyze those rows further.
It also works well against much larger tables. I have tested it out against 25M, 50M, 100M, and 200M+ row tables.
I'm working on an app called Limberly. It focuses on health and ergonomics for sedentary workers - probably most of us here :)
It is scientifically proven[1] that sitting is detrimental to our health, with increased mortality rates. The primary way to reduce the negative effects of sedentary work is to move.
This means doing sessions of resistance training (gym), running, biking, but also taking micro-breaks during work sessions and performing light exercises and stretches.
Research has shown[2] that taking short breaks during work reduces fatigue, and in some cases actually boosts performance.
There are plenty of running and gym apps out there, so Limberly focuses on the last part - helping you take micro-breaks, reminding you to change your posture between sitting and standing, changing which hand holds the mouse (if you're into that) etc.
It is still in early development, so if you'd like to help test and shape the app as we go, please sign up for the waitlist and I'll add you to the testers group. Feel free to also DM me here with any questions or feedback.
Oh, I am also writing a series of articles that explains it more in depth: https://prodzen.dev/articles/building-limberly-part-1-we-re-...
1: https://pmc.ncbi.nlm.nih.gov/articles/PMC10799265/
2: https://journals.plos.org/plosone/article?id=10.1371/journal...
I'm building Deptheos (https://github.com/ryanmitts/deptheos), a photography/macro focus stacking program.
It works on MacOS, built with Swift and Metal. My goal is to make a super fast, and free, focus stacking program. I provided a notarized MacOS DMG for the initial release, but if built yourself, it will run on an M4/M5 series iPad Pro as well.
The core ability I wanted was to support RAW files as inputs, with DNG files as outputs. This is done using either LibRaw, or Adobe DNG Converter (runtime options).
I have been really into macro photography the last couple years, and have been slowly working on trying to build my own program to handle the focus stacking.
Building Alyph — a canvas-based AI workspace.
My problem: my super-long, months-old ChatGPT threads were breaking down. Even typing got slow, and the longer the threads got, the more they hallucinated. I loved Google AI Studio and paid for it, but I was constantly deleting and re-editing the same thread just to try a different angle. And I couldn't run multiple frontier models against the same context or files without copy & paste and tab switching.
I do 99% of my AI work now in Alyph. One board instead of a dozen tabs, branch anywhere, hot-swap models on the same context. Best guess: I'm 3x faster building things. The honest 1% failure case: it's too slow to load for a quick throwaway question.
Hardest technical problem so far: layout algorithms for an infinite canvas.
Pre-revenue, early users. Building self-service billing now.
Looking for a co-founder who's sold B2B software before.
A complete guide on how to travel or live by van in Japan. Additionally trying to turn my passion into a revenue by offering tourists custom handcrafted plans for them to travel.
This is a fun side project as I learn great with email communication, culture differences (as a dev)
I've been on sabbatical (not on leave from anywhere, just decided to take a break from work) for months now, taking some time for myself. Minimal tech stuff until more recently, but now I'm back in the deep end.
The main thing I'm currently working on is a platform for organizing and discovering in-person events. Still not certain about the boundaries for "Phase 1", but I have a bunch of ideas in that space that I've been incubating for a while. One subset of features will be roughly similar to that app you've probably heard of that starts with 'M' and ends with 'p', but hopefully an improvement, at least for the right audience. But wait, there's more. :)
Currently building it; it's not public yet, so no link. Next month. I also have an external deadline around that time.
Thinking about how to grow the userbase is intimidating, but I think it might end up being fun.
I am building Chief of Staff (getcos.ai), an AI assistant for startup CEOs specifically (and executives with 100+ emails a day in general), because neither Superhuman nor Lindy were working for me.
What I've noticed personally and with founders I talked to is that communication and email triage takes a large amount of time each day, but mostly only needs a quick decision or rerouting the right kind of information. But all this mental overhead takes away time from the really important strategic tasks that you should be working on as a founder.
That's where Chief of Staff comes in. Like a real chief of staff, it gives you a head start into the day: checks your calendar, your incoming emails and messages, prepares meeting briefs and drafts responses.
But it goes further than your typical smart inbox: by sharing your strategic goals it prioritizes and makes sure you are working on the highest leverage items. It also helps you manage relationships by tracking communication frequency and sentiment, so you don't miss when a key customer goes cold.
Both Superhuman and Lindy fell short in key areas of the UX for me: Superhuman makes email triaging faster, but you're still the one doing most of the work. Lindy is highly customizable, but you'll spend a ton of time building and tweaking workflows. I wanted a batteries-included approach to get started right away by just connecting Gmail, Slack, and Calendar without any additional configuration.
A key UX decision for me was also that I stay fully in control. Chief of Staff reviews, analyzes and prepares, but I am the one hitting send.
I'm testing it with a very small crowd right now, but want to open it up soon. If you feel that this is an issue you want solved for yourself as well, feel free to reach out to me and I'll get you on the list for the private beta.
If you have strong opinions for or against this approach, I wouldn't mind hearing this either :-)
Koji, a AI Backend Orchestration/Managment/Proxy layer: https://github.com/danielcherubini/koji
I've been running models on my homelab for a bit now, but none of the available options out there was what I wanted. I wanted something that I could command from the CLI, API, or Web, so have an agent go in and do work remotely via SSH or myself via a web interface.
I wanted the ability to know if models have been updated, and if backends (llama.cpp, ik_llama.cpp) have been updated, see what those updates are and choose to update. Also wanted the ability to switch betwen versions of those, so if I felt like there was a regression, or performance issue, I could roll back.
I've also published plugins for OpenCode and Pi so that model discovery is automatic too.
I'm building this mostly for me, as usual.
Cooperation Cube (https://cooperationcube.com/) — A strategic 4-player memory/semi-cooperative board game I designed, played on a rotating 3D cube. Just added a daily puzzle (https://cooperationcube.com/daily) you can play without signing up. Place sticks, complete patterns, and try to beat the day's challenge.
Live Kaiwa (https://livekaiwa.com/) — A real-time Japanese conversation assistant. It listens, transcribes, translates, and suggests responses so you can follow along in conversations you'd otherwise get lost in. I built it because I live in Japan and needed something for the situations where missing a nuance actually matters — PTA meetings, bank appointments, neighborhood councils.
Back in the day, my friends and I loved to rip a few games of Curve Fever 2. The original is gone and the game that took its place has objectionable aesthetic and gameplay tweaks.I've been working on making my dream "curve-like" game that captures the elegance of the original's gameplay while also allowing optional stuff like portals, rocket launchers, custom maps and modes like capture-the-flag. I'm kind of going for that sense of hilarity and semi-competitiveness of e.g. Halo 3 custom maps and modes.
My friends and I have been having a great time playing the initial version, and it's been fun working on some of the more interesting technical aspects like server + browser performance, mapping 2-d game space onto a 3-D visual space, etc. as well as some just-because-I-want to things like a dynamic music system.
Still working on:
- https://github.com/rumca-js/Internet-Places-Database - Internet meta database
- https://github.com/rumca-js/Internet-feeds - list of Internet feeds
- https://github.com/rumca-js/crawler-buddy - crawling framework
- https://github.com/rumca-js/RSS-Link-Database-2026 - link meta from year 2026
- https://github.com/rumca-js/RSS-Link-Database-2025 - link meta from year 2025
Working on IBAI Architect a digital identity strategy focused on bridging the trust gap between AI startups and enterprise buyers. We specialize in acquiring category defining assets for AI safety and infrastructure. The goal is to help founders own their global namespace from day one, especially as the industry shifts toward physical AI and rigorous safety compliance.
Building Braindump AI a simple privacy first voice to text PWA that runs entirely in the browser with zero API costs. Speak and it auto sorts into tasks, ideas and reminders. On the side, thinking a lot about digital identity and naming strategy for AI tools especially how founders can build trust signals from day one as things scale toward enterprise.
Would be interested to hear what others are wrestling with on the branding or infrastructure side
https://statusdrift.com — uptime/cert/domain/dns/tcp monitoring, with status pages, incident management, SLA monitoring and on-call.
I got tired of the state of monitoring and ITSM tools. Most established tools stopped investing years ago. Everything has artificial limits or a credit system. Incident management and status pages are always a separate product. Used ServiceNow On-call quite alot, but it is too slow, and complex to setup simple schedules. Good look with overrides also. Uptime Kuma is modern and great for hoby projects, but lacks other features for smaller teams or agencies. So I built StatusDrift to be the one tool, flat monthly rate, no per-check credits, and a free tier for commercial or hobby use.
Would love to hear what you think.
I got sick of answering the same property management questions over and over again for my rental properties, and am finally building a modern, fully-fledged property management system, AveryIQ - https://averyiq.com
Appfolio, TurboTenant, Innago, etc. are just an extra chore, they don't actually help you do the work. AveryIQ is an extension of your communication channels - it handles everything from email, SMS, and voice calls. The idea is that you could go on vacation for a month and a moderately-sized long term rental portfolio could run itself. You get texted once a week for the important decisions (work order vendor bid approvals, lease signatures, etc.)
I am working on a TUI framework for tokamak https://github.com/cztomsik/tokamak so I can then use it for my local agent harness. The currently separate PoC now lives in codeberg https://codeberg.org/cztomsik/hello-tui
I'm building https://jsonquery.app to query, store, and organize JSONs, with natural language querying so you don't have to learn another query language.
Also building BetterGit: (https://www.satishmaha.com/BetterGit/) A simpler, cross platform Git GUI where all the commonly used actions are right in front of you
And also Crush Depth: A remake (from 13y ago) of a tower defense game for Apple's platforms (iOS, iPadOS, and macOS). Checkout the TestFlight: https://testflight.apple.com/join/gkD5c2U1
I am working on bUniProbe (https://www.crowdsupply.com/bitmerse/buniprobe), an open-source wireless hardware debugger. I was tired of juggling separate USB adapters for SPI, I2C, UART, CAN and GPIO and wiring up external level shifters just to switch between 3.3V and 5V targets.
To fix this, I built a single Wi-Fi connected board that handles it all. It hosts its own web server, so you can monitor signals, read/write data, and toggle hardware pull-ups directly from your browser without installing drivers. I also added a waveform viewer and a REST API for all interfaces, making it easy to automate hardware testing with Python scripts.
Hardware and firmware will be fully open-source. We are currently in pre-launch on Crowd Supply.
I'm making capability security for distributed systems. The primitives and engine are both open source. Primitives: https://github.com/Hessra-Labs/hessra-tokens Engine: https://github.com/Hessra-Labs/hessra-cap
It's built using biscuits and written in rust. I'm really into it. Using capability security as a model makes building things feel like they snap together a lot more naturally. At least for me.
I've also got a blog post describing it in more detail: https://www.hessra.net/blog/what-problem-led-me-to-capabilit...
Making 3d web games with webgl. And wondering if I should go all in on a career switch into digital art and 3d and leave software.
I built a tool that generates custom visualizations to prepare for big moments: http://bevictorious.co/
I'm an avid weekend golfer trying to improve. I noticed a lot of negative self-talk after a poor streak on the course and picked up "How Champions Think" as recommended by r/golf. In a chapter on optimism Bob Rotella shares a story about how, before he won the Master's for the first time, Seve's friend made him a tape of a fake news broadcast of him winning it that he listened to obsessively leading up to the tournament. Victorious lets users generate similar visualizations for any big moment (sports, performances, public speaking, job interviews, hard conversations...).
Git worktrees are awesome but they broke my workflow in a couple ways:
Resuming work. I used to `j <reponame>` then `gco <branchname>`. Now if I do that I get an error about the branch being checked out already in another worktree. I realized the branch names are pretty unique across repos so I made ` jbr <branchname>` that works from anywhere.
Jumping within repo. The other kink was when I wanted to focus on a particular package I’d do `j <subdir>` and it would usually be unique enough to jump to the one in my current checkout. But now I have dozens of concurrent checkouts and have to pick, even though I’m already in the repo. So `jd <subdir>` does like autojump or zoxide but only within the current checkout.
To power those shell functions I made a “where” extension for Git.
https://github.com/turadg/git-where
It’s working out nicely!
I am working on a CLI tool, that deploys managed services on bare VMs, see a live demo here: https://pellepelster.github.io/solidblocks/cloud/index.html
Currently PostgreSQL, Generic Docker services and GarageFs is supported, Backup and disaster recovery included. Aimed at smaller setups that do not need the complexity of the big Hyperscalers.
Working on my own agentic environment that focuses on short lived agents and sandboxing. Agents are short lived and are spawned by the planner you're interacting with. This way models with low temp settings can focus on a single task, don't overflow, and produce actually good and readable code.
Only supports Go, specifically, due to unified formatting, codestyle, testing methodologies, core packages etc. Saves tons of tokens for system prompts to use Go.
Currently working on the summarizer agent and the requirements specifications, because I want to have the specifications check for valid Go syntax (using the upstream parser and ast package).
Warning, here be dragons:
I have been working for several months (maybe half a year) on the IUP fork, a cross-platform UI library with native controls https://github.com/gen2brain/iup-go . I added more drivers/backends, Cocoa, WinUI, Qt, GTK4, FLTK, and EFL. New controls, like Table, Popover, Toggle SWITCH, and modern Tray support. WebBrowser control uses modern browsers; there is a JAVASCRIPT attribute for proper hybrid native/web apps. GLCanvas now uses EGL for both X11 and Wayland, and of course, there is proper Wayland support for all Linux toolkits. I still have a long list of TODO with issues I want to fix, all the docs that need updating, etc. The plan is to post a "Show HN" after I finish with all current tasks.
I have completed by SaaS after 3month. Now working on marketing strategy and directory for launch
Link is below if anyone interested. Catch The Signal: https://catchthesignal.com/
Building FocusedFeed (https://focusedfeed.app) an iOS app that lets you organize your YouTube subscriptions into categories and actually find the creators you care about. YouTube's home feed is basically useless if you're subscribed to 200+ channels. Half of it is Shorts, the other half is stuff the algorithm thinks you want. I just wanted a simple way to say "show me my tech channels" or "show me my cooking channels" without digging through noise.
So I built this, you sign in with Google, it pulls your subscriptions, and you group them however you want. Shorts get filtered out automatically. Working on AI newsletter digests too, so you get a weekly summary of what your favorite channels posted without having to open YouTube at all.
There's a web app(app.focusedfeed.app) but the iOS app is where I spend most of my time now.