alt Hacker NewsAsk HN: What Are You Working On? (June 2025)
What are you working on? Any new ideas which you're thinking about?
Comments
Using Google's GDELT to analyze velocity and sentiment around public-policy/political news. Objectives: develop a taxonomy of news-event types and their behavior; use that taxonomy to test faster/better time to market with responses; ultimately determine which scenarios, if any, can be predicted.
I'm working on a simple app to give a second life to old phones and tablets, turning them into an extra screen with virtual keys (like Touch Portal but free, open source and Linux-first).
Nothing to show yet, still in development, I hope I can share a github link in one or two months.
I am working on a unit-aware arithmetic library for swi-prolog (1) modeled after the c++ mp-units library (2). Turns out prolog is really well suited for this because:
* of its ability to store unit system data as code
* unit conversion is an iterative deepening depth first search
* manipulating symbolic arithmetic is so easy
Unfortunately, it requires users to compile swi-prolog for source because the library is using some unreleased features. If anyone would like to test and report some feedback, I would be truly grateful !
https://tinkerdeck.com/projects/rent-buy-growth
I've been basing one of the biggest financial decisions in my life - whether to buy a house - in large part on NYT/NerdWallet Rent-Buy calculators. But when I dig in, it seems that the model is both extremely sensitive to home/S&P500 growth assumptions, and that their defaults aren't well thought through.
This site is my attempt to organize my thoughts on what reasonable defaults should be, and provides an interactive tool to explore housing and S&P500 growth historical growth rates.
I'd appreciate feedback!
Repo: https://github.com/arajnoha/phodo ultra minimalist web todo app (php+html+css). stores everthing in single JSON, has a day view, allows you only to mark as done, delete and add. list days back and forth + jump to today. When task is added while listing days, its added to that listed day. PHP and CSS are both below 100 LOC.
Im working on simplyfing the code further. I tried really all of the "productivity" stuff to stay organised. Got angry multiple times, went to pen and paper, was OK, but i felt i just need a slight glimps of tech to make it more functional. Something little more than plaintext file, but not much.
Working on https://qrew.cc
When I worked at larger orgs. Reviewing applicants was a very busy task. I would usually get 100-300 applications for the role. And I never trusted HR team to filter out candidates before interviews, so I would go manually through all the candidates. In the world of AI and automatic ATS systems, I have the same problem. I don't trust AI now to filter and rank candidate resumes for me. I wanted something that enhances my process, but does not replace it.
So i've started working on https://qrew.cc, where AI helps you, but keeps you fully in control.
currently in the process of building a small CPG/food lab/accelerator/incubator. I have access to talent, supply chain, logistics network, retail space and a little bit of capital. I want to take and implement the VC model for SaaS startups in Mexico. I'll have a very specific selection process for products we would take on and I'll provide all the infrastructure to get the product onto store shelves in Mexico, USA and hopefully Europe.
I have no idea what I'm doing but I never really have. I've always just figured it out on the way.
I'm one of the creators of a tool called PBRgen, which is an AI-powered online PBR material generator. The idea is that it can generate seamless materials for games and 3D animations etc. We’re still in early beta and really want to shape this with input from artists, so we are looking for artists and creators to help test and give feedback on the tool.
Here’s the link to try it out: https://pbrgen.aixpoly.com/ (limited spots available)
Let us know what works, what doesn’t, or what you wish it did. All feedback is so helpful right now.
I‘m building Sticker, a simple note-taking app for Mac that lets you add markdown based notes to applications, files and other windows. The note is only shown when the connected file, window or app is selected / has focus. Currently, I use it myself to add notes to specific files and projects, e.g. adding a note to my tax folder for 2025 instead of creating a txt file or adding a ToDo to a specific workspace when opened in VS Code. The notes are completely file/markdown based and can be simply synced with other devices. This way, it’s also possible to edit the note outside of Sticker.
I managed to ship my journaling app [0] focused on language learners. Basically, you write your journals in a different language and see corrections.
Gonna focus on marketing and improving the app.
I've been working on my business for 4 years now, sometimes taking extended breaks when I run out of motivation.
Lately, I've noticed that my (beefy) server is always clogged with background jobs that tend to run longer than they used to. It’s started impacting operations, as customers have been complaining about their backups running a bit late.
We're network bound, so I can't just add more compute power (Notion's API has a rate limit of 2700 req/15 mins). I suspect we're being getting rate limited left and right, which is causing these delays.
I am making a few programs that takes everything I've learned from the 6 months writing 90s.dev and turns them into useful applications. One of them I'm going to release within a few weeks, it aims to be a modern dos+qbasic experience. Another one aims to modernize the experience of having a new computer that does almost nothing and you have to program it in assembly to get it to do stuff, except it'll use wasm (see my Ask HN post for details) with wamr+llvm for near-native performance and SDL3 for full GPU capabilities, and it's called hram.dev (H-RAM = hand-rolled assembly machine). That one needs a little more time to bake, so I have to release the qbasic+dos thing first to keep the lights on. Still thinking of a name...
I am writing a database purely to satify my curiosity and to do something fun . https://github.com/edisontrent17/trentdb/tree/main
In the last 3 months, I've analyzed 540 directories and platforms (did backlink analysis with Semrush, checked whether new tools have been added recently, checked traffic, tried submitting on the platform, checked approval time, special requirements etc.) and curated a list of the best 100+ platforms to launch your new product and gain initial backlinks.
I recently launched a free newsletter where I'll be sharing one platform every day with pro tips based on my experience for the next 100 days.
Check it out here: https://topsaasdirectories.beehiiv.com/subscribe
"ROWM" (read once, write many) robust file copy program that utilizes checksum sidecar files. (1) Read the data, compute a checksum as a 'source of truth' and store it as $filename.checksum_type, and then (2) write the data (and the computed checksum) to N destinations simultaneously, and (3) Compute the checksum of the file on each destination and compare it against the source of truth
Could use "tee" to limit the reading to just one instance but I would like to try Python.
Hoping to write the core of it as an open-source hobby project to learn Python multithreading and then extend it for the actual problem I need to solve at work through the use of config files.
Mobile app that lets you continue coding while you’re away from your computer.
The goal is to be a full mobile IDE that lets you use Claude Code, Gemini CLI, and other agentic code editors.
Has mobile-native file browsing and git integration.
Working on a Professional Upskilling AI Coach: https://www.socratify.com/
We're headed into an era of massive white-collar reskilling.
How you think > What you know.
Critical Thinking skills will be the most important skills as we AI expands throughout the economy and we're surrounding by LLMs that are highly fluent
Socratify is a Critical Thinking Coach that sharpens How You Think and Speak by Debating AI
It proposes interesting questions (currently business related) that you debate in 2 min conversation and get feedback on how you think and speak
Right now its most helpful if you're interviewing for a job or aiming for a promotion in a business related profession
Created a dashboard to help people figure out what renewable energy solutions they could use for their homes in the U.K. https://renewable-home.verdient.co.uk/
Created a game to learn navigational marks in the Solent https://guess-the-mark.verdient.co.uk/
Putting together the landing page for my software business https://verdient.co.uk/
I’m also putting together an analysis of warhammer 40k games and applying operational research techniques to it.
Working on an open source language learning app. It does listening/speaking drills with spaced repetition.
It’s like Anki but for speaking and an LLM grades your response.
I'm working on Tennis Scorigami - a data viz and tennis centric project somewhat similar to NFL's scorigami, but with a little bit more (if i do say so) interesting visualizations / new ways to look at the data.
From a technical side, I've processed around 325k+ matches. Right now, only main ATP / WTA matches (no challengers, no doubles, no mixed) sadly. I'm working on expanding that, improving our infra layout, exposing a public facing API, collecting the data on my own, and most importantly live score ingestion (especially given the fact that Wimbledon is starting tomorrow).
Feedback on the app through Canny / joining the Discord / following the Twitter / or any and all of the above would be much appreciated.
Working on a platform to create virtual personas for social media and educational content. The personas have a consistent personality, looks and voice. Content can be scheduled on social media channels. It is still very much work in progress but already live as a generic content creation platform. https://postcrest.com
Also working on an email communication assistant https://merel.ai creates draft responses for gmail and outlook based on your company data, email history, website content and extensive organisation settings. Still work in progress as well.
Big update to my micro-saas https://testingbee.io!
TestingBee is a way for startups to get part-time QA for their product's critical flows.
I've been working at startups for the last four years and I've consistently been on teams struggling to balance launching quickly versus keeping our product working. We've never had success creating a substantial test suite because our product is changing too fast and engineers are too overloaded.
I built testingbee as the solution. It lets you write your app's flows in plain english and the bot I created will execute those flows in your app as a user would. This triggers on every push to make sure every release keeps your product working :)
New “AI in a box” product, can run the big models I.e. DeepSeek-R1-0528 etc. comparatively cheap, fast and just works. Our build partner is big on sustainability, considering a return to upgrade option.
Likely will do a prosumer SKU, will be faster and cheaper than the Mac Studio equivalent.
I've been working on and off on a client-side* SERP rank tracker: https://serpowl.com/
I wanted a simpler alternative to the self-hosted SerpBear tool that I could use and share, so this is the result.
It uses SerpApi (where I work) as the data source for what actually executes the SERP scraping because it's much too complex to have purely client-side, but 100% of the rank tracking portion is client-side.
It's not fully complete and there's definitely rough edges with it, but because of the data source, it supports a large number of search engines right off the bat.
Some major updates to https://weather-sense.leftium.com
Play spot-the-difference with the old screenshot: https://github.com/Leftium/weather-sense#weathersense
- At least five major changes!
- Or look at the commit history ;)
---
I'm designing a game that:
- is simple to play. (just log in and check-in with your geolocation. Optionally add a short message)
- helps people stay connected. (You can view friends/family on the globe with some mild competition/cooperation)
- Right now, I'm trying to figure out something compelling to "collect." Cities/states, weather conditions, letters, numbers, words, etc... I think it should be tangible.
Ephemeral, client-side encrypted sharing of files, text, html, and forms.
Just prototyping at the moment, but the goal is to allow users to not only share files (even big ones) but also forms, like Google forms, but encrypted and one time only (read once).
The use case I have in mind is allowing businesses to create GDPR forms (with private info, consent, etc), share unique urls with specific customers, and once the data is received by the business delete it from the server.
This could be useful to businesses that don't have a customer-facing portal, but have to deal with PII and the customer needs to consent and verify the data and what it's used for.
The data is encrypted client side (web crypto) and the password either shared in the url (in the hash fragment, also encrypted by a key stored on the server) or by other means (eg. could be the recipient's dob or id number or some other previously shared or known value).
Still trying to figure out the details, use cases, business value but the core backend is done so is the client-side crypto stuff. I managed to get chunked AES-GCM working so that it doesn't load the whole file in memory in order to encrypt it, it does that in chunks of let's say 2MB. Chrome also has chunked requests (in addition to responses) for sending the file to the server, but would probably need to come up with some other mechanism to get that working on other browsers (like send the chunks in multiple requests and append to a single file on the server, but that adds more complexity so I'm still working it out).
v3 of https://stacktape.com
Stacktape is a PaaS that deploy to user's own AWS account.
v3 adds many new features, but namely the ability to generate IaC config directly from code, by analyzing the user's repository (both deterministically and using multiple AI techniques).
For example, if it assumes your application is a Web API that uses Postgres and Redis, it will create a Stacktape IaC config that deploys Fargate container, load balancer, Aurora Serverless v2 Postgres and Elasticache Redis (behind the scenes it will also configure things like networking, VPC, security groups, IAM, etc.)
Launching this weekend.
Just like another poster, I'm also building a website of daily puzzles, finally at the point where it's mostly finished and I'm not completely ashamed of it - https://dailyplay.club
https://weexpire.org - An opensource tool for creating emergency notes that can be read by your trusted contacts only after your death or if you are seriously injured.
I have been managing Claude to work on a rational math library in JavaScript: https://calc.ratmath.com
I am particularly enjoying the Stern-Brocot tree exploration: https://calc.ratmath.com/stern-brocot.html#0_1 I hope people will find it to be a nice way of understanding good rational approximations and how they tie into continued fractions and mediants. A nice exercise is to type x^2 in the expression box and go down the path to always advance towards x^2 being 2. This gives the continued fraction representation of the square root of 2.
We host a static site service where users can manage their sites via ssh (https://pgs.sh). Previously we used minio for object storage but have become frustrated by its perf issues on smaller VMs, don't need the distributed features, and wanted something a little lighter weight. We initially thought Garage could check most of our boxes but very quickly discovered perf issues there as well.
So we decided to build out our own filesystem adapter and recently deployed it. It's pretty exciting to have our own solution that does exactly what we need and appears significantly faster.
It makes us want to open source pgs.sh because it has fewer dependencies in order to deploy.
I'm working on an app that I use to track my sets/reps. I've added in a trends screen (both for muscle groups and for individual exercises) to check on if I'm actually progressively overloading on a consistent basis. I'm currently adding an AI wrapper around the data to figure out better insights about certain exercises, what I'm missing, etc.
Writing a go binary to act as a wrapper around ripgrep and fzf. Can be done in many ways but I wanted a simple binary that I can invoke from lf or the command line to search, so that I'm using the same keystrokes to search, inside or outside of editor.
I made Peekly[0] because I was tired of feeling FOMO about all the stuff happening in AI, dev tools, indie hacking, etc. I couldn’t keep up with blogs, feeds, and newsletters — it was too much.
Peekly pulls from high-quality sources using LLM + retrieval, then sends you a regular digest with just the most relevant content according to your interests. You can even give it custom prompts to control what it finds and how it summarizes — super useful if you want a particular angle on a topic.
YC folks can use code YC256 for an extra free month (on top of the 14-day trial). Would love to hear what you think!
Working on https://www.space4ads.com - an Out of Home Advertising Marketplace that connects advertisers with underutilized commercial spaces. Think Airbnb, but for ad placements. We're targeting vacant storefronts, construction walls, cargo vans, box trucks, 18-wheelers, and basically any commercial space with foot traffic or vehicle visibility that's not being monetized for advertising.
Repo: https://github.com/specfy/getstack
I'm building this website to track technology trends and usage across the most popular GitHub repositories. I parse 35K repos every week and find tech inside, aggregate it in Clickhouse, and show a summary on the website.
It was a good opportunity for me to finally learn more about Clickhouse, also trying to fully self-host on a VPS, which has its own challenges, especially regarding hosting frontend with SSR
Projects
- https://github.com/rumca-js/Internet-Places-Database - Database of Internet domains, links
- https://github.com/rumca-js/Django-link-archive - RSS client, web crawler
- https://github.com/rumca-js/crawler-buddy - web scraper, web crawler, with JSON interface
A project is like a pet. You cannot just "stop" caring about it. If it lives, then you have to look after it
This goes against the grain - but I’m starting to reimagine enterprise & manufacturing workflows using AI.
Very good leadership support, small (but great) team, huge mandate and uncertainties - and it’s quite exciting.
We are also hiring - AI+Full Stack https://cenovus.wd3.myworkdayjobs.com/en-US/careers/job/Seni...
I'm building a site that tracks the price of power tool batteries (currently Milwaukee and Dewalt). It's been an itch I've wanted to scratch for a while so I would know when the battery was a good price vs. the Ah rating. The prices are scraped nearly daily and plotted on a price-per-ah vs. total Ah chart.
I'm working on a tool to remix/manage my playlists, that's agnostic from the different music platforms.
I used to have an integration in Spotify, that automatically copied my "Discover weekly" playlist into an archive. Over time, it grew close to 10000 songs. It also started to get polluted by ambient sound and kids songs when my daughter was born.
I wanted to clean it up but as far as I could tell, the only way was to do it manually, song by song. I'd want to have something more powerful, that would easily let me rearrange/split/curate my playlists based on any arbitrary constraint.
I've been working on AltStack.jp [1], a curated directory of Japanese digital services (think cloud hosting, registrars, email providers, etc.), all made and operated in Japan. It’s for anyone in Japan looking to reduce reliance on foreign (especially US-based) platforms, inspired by projects like European-Alternatives.eu.
The site itself is built with Astro, content is written in Markdown. It's still very much a work in progress: the design’s evolving, search isn’t done yet, and I’ve only scratched the surface with a handful of categories out of the dozens I have planned.
Still working on Orbiter[1] but with some big changes.
Orbiter started as static website and web app hosting, but we recently added what we call server functions (what you call serverless functions). So now you can build full stack apps. Our preferred development stack to support is the new bhvr[2] stack.
Building the Knowii community, a safe haven for people who care about learning and growing: https://store.dsebastien.net/l/knowii
I'm working on Internal Developer Platform for private clouds. Kind of like private Heroku. It works standalone and installed fully automatic. With primary focus on Low Operations and Self-service where app developers can focus on delivering real business value instead of boilerplate tasks or waiting for other teams to plan and execute standard tasks.
We originally started supporting Low-code solution called Mendix. Now we support any type of web app that can be packaged as an OCI image.
You can read or try it at: https://low-ops.com
It's kind of boring but I'm learning k8s and argo-cd to figure out if I can do feature-branch deployment to a cluster.
like, it would be very cool to do something like have your feature branch be deployed to a separate pod in dev cluster, and have an ingress rule set up so that it points to that pod only.
So if your dev environment usually points to <some-app>.dev.example.com,
Deploy your feature branch to a dev cluster, but on a different pod. Then have it reachable to <some-app>.feature-branch-1.dev.example.com without touching main.
I think it's a neat idea and I'm sure it should be possible if I configure some istio settings.
It's all new thing and it's fun to have a direction towards learning
Repos: https://github.com/romshark/tik https://github.com/romshark/toki
I'm trying to make i18n easier, integrate it better with CI/CD and automate it more with LLMs (for now in Go, second priority is TypeScript and other languages later).
For this I had to develop a completely new approach and subsequently a specification for the "textual internationalization key" (TIK) which are programmatically translatable to ICU MF.
Toki is the first TIK processor implementation for Go.
A surreal open world mystery game in the “Outer Wilds-like” genre - sometimes referred to as “Metroidbrainia”
Just made the first devlog video: https://youtu.be/CFgDlAthcuA
https://hobzcalvin.github.io/blumon/editor
Node based visual editor for 2D LED patterns over BLE. Web/iOS/Android app to ESP32, works with most addressable LEDs. It’s like TouchDesigner x WLED x PixelBlaze, but Bluetooth so you don’t need annoying wifi setup. And hopefully you can make much more interesting patterns without touching any code.
Eventually the ESP32 devices will save all the patterns they’ve seen and share them with apps that connect to them. So there’s a pattern ecosystem, like Electric Sheep.
Still rough and in progress (and constantly deploying so it may break for you )
I'm bootstrapping my predictive database startup https://aito.ai/ :-)
More specifically, I have worked on the demo https://github.com/AitoDotAI/aito-demo to make use cases visual and well described. E.g. smart search use case is here https://github.com/AitoDotAI/aito-demo/blob/main/docs/use-ca...
Claude Code is doing absolute wonders on setting things up. One has to just check out for hallucinations and made-up stuff in any written content.