alt Hacker NewsAsk HN: What are you working on? (August 2026)
What are you working on? What have you been curious about lately?
Comments
An agent harness that is kind of like airflow for agent tasks.
I thought open claw was cool but could never use something so insecure.
Each run is in a container. One network bridge out so everything from tool calls to model communication goes over the bridge.
Each run has a capability token and the host checks grants for the tools attempting to be called.
Core super thin and everything from credentials handling to model communication is a package.
Finally got it alpha and have some useful jobs set up like it triggering jobs on an alert channel in slack. Debugging it with tools like GitHub kube and our work mcp and the dropping a comment on every alert with the cause.
Still closed source but hoping to change that after some more testing.
I've been resuming work on a Zig library for developing N64 homebrew¹. Previous milestone was getting a framebuffer onto the screen, and after achieving that I took an overly-long break.
Current goal, which I've been working on for a bit more than a week now, is to add support for controlling the RDP (the part of the N64's GPU that actually does all the rasterization), such that programs can actually draw triangles. I'd already implemented the control registers, so the current work is defining all the RDP commands, along with documenting what they do and how to use them — which means that I need to fully understand what they do and how to use them, and that's been a fun rabbit hole to explore.
I'm currently about halfway to two-thirds of the way through defining packed structs for all the RDP commands, then the plan is to build a CommandList struct with methods to build RDP command lists and validate that they make sense (make sure all the triangle/rectangle rendering commands have all their subcommands in the right order, make sure you ain't calling them without there being a framebuffer defined as the rendering target, make sure the needed synchronization commands are in place where needed, etc.). Then that should hopefully be enough for Zig64 to draw its first triangle :)
----
I have been working on a static site generator. I've used Hugo and Jekyll before but everytime there is something you have to learn or get used to and they never map 1 to 1 to the flat structure of a simple blog for example.
It's a few classes of C# right now that parses a folder and auto generates the navigation, links, titles and summary off of the folder/file structure for the 1st level and then subsequent depth is mapped to individual items in the list view not on the nav bar (although I did that too, it looked weird with higher depth). So if you have a folder with 3 files and 2 additional folder you get 5 links in the nav bar, 3 content pages and 2 list pages. No other setup is needed. No llm written code, but I used deepseek v4 flash to ask various questions and get samples of some things.
I forgot just how much depth there is to exploring a simple problem. Loving it!
Photo upscaling and inpainting apps, both running in the browser https://crisp.photos https://wipe.photos
Just a little personal tool: Shipment tracking I call Lighthouse:
- All our shipments in one UI (instead of searching trough our mail boxes or going to several merchant apps/sites)
- LLM processes update emails, but everything can be manually adjusted (or created from scratch)
- We have a bunch of destinations we order to (work, home, different pick-up spots), this lets me see where I have to pick up and with which pick-up code
- Real-time sync in case we update something, the other person sees it immediately
- PWA, with animations, feels as good as native on mobile
- $0 running cost (I only paid for AI sub when vibing this)
Desktop: https://imgur.com/a/HD9mCIr Mobile: https://imgur.com/a/EAuJiqH
I'm working on a platform that aggregates a business's operational and financial data and builds a causal model on top, deriving target metrics like FCF or EBITDA. With that model we can forecast different scenarios — what happens if we open a new shop, raise prices, take a loan. We also generate reports: weekly overviews, anomaly detection, deviations from targets. All of this is essentially a harness for an AI agent that customizes the platform for a specific business and integrates its data sources. Large companies have dedicated teams and sophisticated tooling for this; my bet is that small and medium businesses would benefit just as much, and with recent progress in AI it's finally possible to deliver it tailored to each case at very low cost.
My primary target right now is e-commerce due to rich data streams, however if you think your business would benefit from this tool, I would be happy to connect and discuss it.
Working on https://radius.to/ - a Meetup.com alternative of sorts - with fairer organiser pricing for smaller groups.
I posted a Show HN [1] here a while back, got tons of great feedback, and have been slowly improving it since, with no real marketing. Another Show HN is probably (definitely) overdue!
Current focus is shipping a "Meet people doing X" feature - i.e. "I'm going for a cycle today in Richmond Park - who wants to join" (independent from groups) so I'm excited for that. There are some UI examples of how that will look on the landing page.
I'm still working on https://openaltfinder.com; a directory to help people discover open-source projects. Added quite a bit of new features over the last month, including a new weekly newsletter, and a way for people to find beginner friendly issues: https://openaltfinder.com/contribute
It's been tough, for some reason both Google and Bing completely erased my site from their indexes.
Another thing I'm working on: https://getpinnd.com - A small social network for map makers.
I got laid off recently and finally decided to dedicate some time to learn Rust, I started to work on an an idea I've head for a while, to create a game server to play Thai chess inspired by lichess.org, in fact, I plan to use several of their assets.
I'll use Rust for the backend even though it's probably an over kill, but it's the best way to learn for me, doing something with the language. I have some of the front end move logic and board done, I plan to start working on the API soon, after I finish a Rust book I'm reading. Here's some of the code if anybody is interested: https://github.com/thaichess-org.
https://nommer.ai turns recipes into 2-player mode so you can cook together
People seem to love the concept so far and are starting to sign up, now just need to figure out how to get them to actually use it
I've been working on shard balancing for Playwright Test. The idea is to use the new API released in Playwright v1.62 together with historical test durations to balance tests across shards.
I ended up designing a specialized variant of LPT that works well for Playwright test suites. It's now used in WordPress's Gutenberg project, where it cut test runtime from 35 minutes to 23.
The algorithm is called SALT. Here's the detailed write-up if you're interested in job scheduling:
Mu - Tools for Agents. An MCP server and web app that provides real time access to news, markets, weather, video, places, search, etc. It's sort of extension of a previous idea I had to put all the key tools and services behind a single api and access token. I don't know that it's quite there yet. The idea might change. Still looking for feedback.
https://micro.mu https://github.com/micro/mu
Also have an agent harness and service framework in Go
I have been building a PDF reader with dark mode that supports image properly with vimimum navigations. Been using it for the past few months and really happy with the result. Finally got around to publishing it here: https://github.com/moomou/xeil
It’s not as cool as most of the other projects, but I have a little app (Ato on AppStore - https://apps.apple.com/app/ato/id6757285141) for basically tracking all walks/bike rides I’ve ever been and visualize it on a single map. I’ve had the app for a couple of years, but redid it with “hexes”, instead of just routes. It has a bit of coop, competition and ranking feeling to it nowadays as well.
I started the original project as an excuse to learn SwiftUI, then been using it as an excuse to test all the frontier models and etc. to see how much better we’re getting at AI-coding. Kinda fun to see a couple dozen weekly users semi-competing with each other. Obviously nothing big, but the feeling you get seeing someone else use your app is very nice.
I am working on a CLI tool that will build a static web page for a small brick and mortar store. The software is FOSS and could be useful for the small store that has a lot of foot traffic, but is under served by many of the big tech offerings. The idea is that foot traffic can be increased by using a QR code placed on a banner outside the store directing them to the website that can display sales and types of items that the store offers. Pleases check out the build log for details of my progress.
If you’re in the Minneapolis / Saint Paul metro area and care about trees, you might like my recent project: https://msptrees.com
On a walk and wonder what species is shading you? Curious if the city thinks your boulevard tree is as sickly as you think it is? Find your spot on MSP Trees.
It's been fun working with some public city data and thinking about how to normalize information across different municipalities. Community doesn't end at city boundaries, but municipal GIS data systems sure do. I'm hoping to build tools like this for some more areas if I can track the data down.
Recently bought a Sovol SV08 printer. It comes with open source binaries, but there is no source code. Turns out they are a known to be violating the GPL. Looks like it can be rebuilt from other sources, so I have started to put together a clean implementation using Yocto. It will be possible to add support for similar 3D printers.
I bought that printer in order to print parts for two other projects: a baby monitor (security camera) and a hat-mounted teleprompter monocle. Both using COTS part and running Yocto-based open source images. All affordable products in both of those sectors are completely untrustworthy or otherwise unsuitable for my requirements, since one of which is total replacement of the firmware with my own custom versions.
In my day job, I am a principle embedded engineer consultant working on distributed embedded systems, so all of the above software will be relatively easy.
The mechanical work will be the most challenging, as I presently only using OpenSCAD. I would entertain suggestions for alternative open source CAD solutions, but I hope to recruit a ME (or two!) to help drive that portion of the work to completion.
I'm working on Spacecap: https://github.com/mgerb/spacecap
It's a screen recording tool for Linux. The goal is to make it as fast and efficient as possible, while also being simple to use and install. It's just a hobby project that I am having fun working on. Here are some notable things about it:
- Written in Zig
- Uses Vulkan Video for encoding
- UI built with SDL3 and imgui
- Features recording, replay buffer, and screenshots
I'm working on writing up pages for projecthammer.org, the new website where I'll host my database of Canadian grocery prices. It's getting interest from more academics and needs a more formal home.
I've been curious about ways in which politicians who are in the minority/out of power can still make progress on issues that matter to them.
I'm working on a tool for visualizing biological pathways, pulling data from the Reactome knowledge graph for now. Trying out different graph visualization algorithms, it's an interesting problem space as they are all approximations to solving a problem that I believe is NP-hard-- arranging a graph such that all vertices are distinct and you have the global minimum of overlapping edges. Next step is implementing a novel algorithm that I found described in a dissertation that doesn't share its code. Unfortunately it's not deployed yet, but if anyone seems interested I will be sure to share it on HN when it's in an acceptable state.
A visually grounded AI tutoring application: https://knowable.ca Been working full-time on it for a few months, I'm trying to make the concept of an AI tutor better than just a chatbox.
Building a modern typing application.
I believe typing is probably one of the most important skills in 21st century for knowledge workers and yet one of the most neglected. Hoping to change that.
Anyone can learn to type fast, without having to think about or look down at the keyboard. We are building the best tool to help people get there quickly
I'm working on Eland (Extensible Layered Algorithm for Network Drawing), a graph layout library strongly inspired by Elk, but a complete reimplementation.
The reason I need a complete reimplementation is that existing graph layout libraries tend to be black boxes that you can't really fiddle with. If you want to tweak part of the algorithm, you have to fork it or write your own. With Eland, you can replace individual steps with your own implementation while keeping the rest of the algorithm unchanged. To ensure validity of the various steps, I've introduced contracts where every step specifies what properties it requires from the graph, which properties it promises to add, and which it invalidates, which I think is pretty cool.
The reason I need all of this is that I'm working on a system to display and browse through data lineage graphs. I've noticed through experience that existing graph layout libraries don't do quite what I need, so I ended up building my own.
I've been working on a general repository linter. The idea is to declaratively define a set of rules/conventions for your repo, (for things that language-specific linters don't cover), covering things from directory structure, required files, file staleness/freshness, file size, rules for binary files, rules around use of invisible Unicode, etc. - which can be checked deterministically. Many/most large repos have a set of hand-authored scripts for doing these kinds of checks, the tool I'm building basically packages these in a fast, reusable and extensible tool, with some niceties added.
I spent few months to buil the alternative to Wireshark. Native macOS app, built with Cocoa (Sorry SwiftUI), on top of libwireshark, so you can see the packet details like Wireshark. Support TCP, UDP, DNS, HTTP2, ...
For the last couple of months I was working daily, and intensely, on Meltdown. The project started in 2024, but I retook it recently and improved/extended it a lot.
This is an AI Harness + Platform, made in Python and Tkinter.
It has a lot of advanced features, arguments, commands, shortcuts.
Lots of widgets and the markdown parser were done from scratch.
It has a powerful API to create conversational sessions interactively.
You can learn about it on a blog post I made:
https://github.com/madprops/blog/blob/main/docs/meltdown/mel...
I am building a new modern Terminal browser. A text based browser for the terminal. Full privacy, features. Ultra fast. For those who love to work in terminal but somehow are dissapointed with Lynx.
* Native and small. * Private by default. * Terminal-safe. * Agent-ready. A built-in MCP server exposes read-only tools (browser_open, browser_read, browser_list_links). * Configurable.
[0] [code] https://github.com/meerita/puma-browser
I am still far away of 1.0, but I've been using it every day since. It's open source, you can compile it and use it as you want.
Every week I try to finish more features.
I left my job a few months ago and have been building https://oristudio.pages.dev. It’s an origami design workspace that runs in the browser. Super niche but I’ve been really loving just working on a passion project I can use for my hobby.
Two open-source initiatives:
1. https://preseason.ai Basically a benchmarking platform that measures how the top LLMs recommend devtools when prompted to build different types of web apps.
2. https://widen.dev A native Mac desktop Postgres GUI with text-to-SQL. You can even use the LLM that already comes installed in your MacOS.
I wanted to know if I could fully describe an online survey using a simple, readable, parseable language.
There are numerous UI form builders. And Xlsform for ODK and KoboToolbox but nothing I could find that is easy for humans and machines to grok. So I wrote a DSL in Ruby to see if it could be done. The language is mostly described at https://srp.solutions and some example usage and implementation at https://srpsurveys.com/
Working on https://letterspractice.com
This is a high efficiency, narrowly-scoped, low screen-time early literacy app for families with kids aged 2+.
I'm working on my typed, fast data frame language Ibex: https://github.com/bobjansen/Ibex
It is usable stand alone but also through an R package:
https://bobjansen.net/faster-than-data-table-introducing-ibe...
It performs better than data.table in my benchmarking on a wide array of tasks and is competitive with Polars on low core counts.
I'm working on a totally vibe coded clone of Elite - https://github.com/atomic14/harmless - so far it's pretty playable.
You can try it out here: https://harmless.atomic14.com
I'm planning on doing it Open Vibe Code. Add an issue and I'll set Claude on it.
PRs will be rejected out of hand. I don't want your code. I want your words.
It's just a bit of fun to see what's possible.
Research on an AI super agent that manages memory, knowledge, skills, tools, other agents, channels, management and orchestration, and auto update its own code at the end of every session if needed
Previous agents developed:
I am building this:
It is the most comprehensive collection of benchmarks of the safety, ethics, and values of LLMs. I also aggregate sparse safety benchmark scores into a global ranking by optimizing the global ranking to maximize correlation with individual benchmarks.
I'm working on a family organizer app Mana - The AI-Connected Home App for Households & Trusted Circles https://share.google/dCKfODeN7c1QhHhft
A lot of Claude plugins.
https://github.com/mrieck/demoday-claude-plugin - create demo videos for your projects using Claude
https://github.com/mrieck/overboard-plugin - CTO assistant, PM helper Claude plugin
https://plugmyplugin.com - Weekly voting contest and directory for Claude Plugins similar to Devhunt/ProductHunt
I'm slowly working on a spacecraft design webapp: https://vesperastro.com. Think PCPartPicker for spacecraft.
You can build out and trade hardware in the master equipment list and immediately visualize how any changes impact mass, power, propellant, and link margins, or break power or data interface compatibility. You can also play around with orbit and operating modes to optimize things like orbit average power.
I'm hoping it can be useful in rapidly iterating early stage spacecraft designs and weighing system-level trade studies across CONOPS and hardware. Its still very much a work in progress so I'm always open to feedback.
Helping a friend build a midi controller replica of a legendary vintage synthesizer (Yamaha CS80)
For the last 6 months I've been working on a local-first replacement for accounting and business management software. I just finished building a modular system that I'm pretty proud of, which lets users develop their own modules and import them into a running instance without having to fork the repo.
I'm working on a browser IDLE game made with Go. I've never touched Go before. The game core it's 80% done and it was pretty easy to code but I have no idea what I'm doing in with the transport/http layer.
At this rate I'll have it finished by the end of the next decade.
I am building my "dream tool" too, a cross-device roaming profile-experience context-focused UI. Many, many iterations that go back maybe 15y, even built a custom linux distro (iolinux) back in the day(ro root, containerized user-spac with composer-like app handling). Always worked on side-projects, consulting gigs or one-off implementations, often times across multiple customer-devices. Switching context is damn expensive, hence tried to build a tool to help with that. Always failed on the UI parts - native linux tooling was just awful. What do you specifically mean by "context" you ask? Well, you get a call, switch to work on `work://customer-foo/project-bar/task-baz`, every bound device / tool / application loads data related to that task (your filesystem mount shows relevant files, browser loads relevant tabs, obsidian relevant notes etc). Building my own small db engine to accommodate that use-case, getting more fun working on it every day
Working on https://localhero.ai, automating i18n translations for product teams (runs as a GitHub Action, translates new strings on PRs matching your brand voice and glossary).
Shipped "bring your own translations" feature this month, built together with a customer who wanted Localhero for the workflow and review layer but writes some translations themselves. Otherwise smaller improvements and marketing. Better Lingui support, for example: projects using explicit IDs keep the source text in the source locale's msgstr instead of the msgid, which our CLI got wrong before. Also wrote some blog posts, a comparison of GitHub Actions for i18n and a guide to auto-translating gettext .po files in CI (Django, Phoenix, LinguiJS). Content is slow going but it seems to be how people find us via perplixity/gemini/chatgpt.
On the side, https://infrabase.ai (AI infrastructure tools directory) grew ~16% MoM. I leaned further into the agent angle: added a public query API plus llms.txt/agents.md/etc so agents can query the directory directly, and exported the full dataset to a public GitHub repo. Starting to get som traction and people wanting collaborations.
Building the GoodReads of fanfiction and web fiction. Currently growing really quickly so dealing with scaling pains :/
A few things actually.
I’ve been working up reimplementations of Nikoli puzzles to host on my personal blog, the goal is to work through the maths and then explain how it’s relevant or helps create a fun games. Nikoli is the magazine that helped create sudoku and made it famous - they have a lot of other pen and paper puzzles that are really fun to play.
I’ve also got two other open source tools I’ve built to help me work daily:
https://tajd.github.io/cofferdam/ which is a rust based cli tool I wrote to parse the ast of codebase and enforce architectural preferences - eg don’t reimplement design system for components down to DRY
https://tajd.github.io/projektor/ Is an ai native project management tool where I can plan work on the train and then have my agents pick it up autonomously. Bootstrapped it quickly but then have been incrementally adding more functionality to it. It’s like self hosted jira on cloudflare but much faster and has more agentic development tools. I added tooling around being able to autonomously work through epics, manage quality and fix identified bugs - https://tajd.github.io/projektor/agents/playbooks/epic-goal/. The goal being deliver useful features and then to ultimately surface the number of bugs identified by agents and users so that I can see the health of the project and figure out if I need to write more code myself or improve feature descriptions to reduce discovered bugs. I don’t think this is new at all but it’s reliable and done well, I hope.
Enjoying retirement. Amazed how fast I was able to just stop thinking about software. It's very liberating.
I am working on Cangjie Workbook. I am trying to create the best in class tool for learning Cangjie. It's a Chinese input method based on decomposing characters into sub parts (not necessarily related to how they are written). It has a high learning curve, but it is a lot of fun to type with and helps you remember characters. It is a tool that I use myself. I already have a couple of users that have been giving me feedback. You can use a subset for free and without any login credentials. The complete product is available for a small monthly fee. Any feedback is welcome!
Akariq has taken most of my time [0]. I sell travel eSIM plans for 190+ countries, no app or account needed! Since last time, many HN readers have tried it with very positive reviews. I reply to each and every customer who reach out to me and also implemented some feedback they gave.
On July 24 2026, all major eSIM providers including Akariq faced disruptions. However, Akariq was able to serve all customers even during the downtime either by switching to backup resources or issuing replacement eSIMs free of cost. Some customers of major providers like Airalo experienced serious down time. I wrote more about the incident here and what I am striving to be better in terms of automation etc [1].
One review from a customer in China [2] made all my effort worth it. They said that they were able to use Akariq to access Gmail, WhatsApp etc around the GFW. It even worked smoothly in areas like Beihai and Weizhou island apart from the cities. Tethering with their laptop was helpful too.
[1] https://akariq.com/en/blog/akariq-backup-esim-service-interr...
An overhaul of my original PlayStation bare-metal programming tutorial repo: https://github.com/spicyjpeg/ps1-bare-metal
It was never really meant to be an SDK of sorts that other PS1 homebrew projects would pull as a submodule and depend on, but since it more or less became one I am likely going to split off the examples from the "core" (headers, CMake build scripts and tools). I have also been asked to add more examples, particularly on how to play sounds through the SPU and access the CD-ROM drive, and may cover I/O on some PS1-based arcade systems in the future.
A grid-based flowchart editor designed for speed. It’s optimized for the keyboard with a spreadsheet-like selection and movement mechanism.
I started working on it because I was tired of dragging boxes/arrows around in Miro while making dependency graphs for planning projects.
https://knotend.com/
A while back I did a Show HN about it, and have since rewrote and changed the concept a bit to be easier to use.