Ask HN: What are you working on? (April 2025)

AI bookkeeping for solopreneurs.

As a freelance web dev who also has an Airbnb side hustle, I got tired of expensive bookkeeping for a few transactions per month. I tried DIY, but my time is worth more than that.

Most importantly, both pros and DIY got subtle things wrong and caused me to miss out on thousands of dollars in deductions and credits.

So I’m making an AI bookkeeping chatbot that will handle all that for me. The aim is full automation while surfacing tax deduction and credit opportunities throughout the year. Like wouldn’t it be awesome to just have the research and tax credit or do home office deductions with zero effort?

At the end of the year, Kumbara puts together a series of financial reports that you plug into your tax software or hand to your CPA.

Working hand-in-hand with CPAs and some platform partners on this. Would love to hear from other solopreneurs or engineers who want to help build the future of financial freedom.

https://www.kumbara.money

I've always been into MUDs since the mid-90s and every now and then I get this idea to write one from scratch (again) and I get pretty far and then I lose interest.

The new one is browser based.

The mud engine runs in a web worker and takes advantage of some of the modern web tricks to do stuff. For instance, data files (think: area files that don't change often) can be stored remotely and then cached with a service worker. This allows the MUD to run offline. But that's only fun if you're playing solo.

IO between the UI and worker is handled by message passing.

Multiplayer is handled by the MUD opening an outbound connection (probably websocket) to a connection collector host. Other players would then connect to that host and have their IO routed appropriately. The host can even be smarter: it could be a specialized Discord client, allowing users to play from there. Firebase may also be involved. I don't know.

The important bit is that this is still basically message passing, so the engine won't need to know the difference between the local user and a remote user.

The MUD database would be an IndexedDB. Probably. I haven't thought as much about that yet.

I am sure all of this is theoretically possible, at least.

It tried out vibe coding this weekend. I've been using AI to get smaller examples and ask questions and its been great but past attempts to have it do everything for me have produced code that still needed a lot of changes.

I hit an OpenSearch bug this week where you can't get any browser based requests to work. Its due to zstd becoming a standard part of Accept-Encoding and OpenSearch not correctly supporting it so I wanted to install a browser plugin that modified the browser HTTP request headers to my servers.

I don't know about everyone else but I love that browser plugins are possible but I hate having to find them. Its mostly due to never knowing if you can trust a plugin and even if you find one, you have to worry about it being bought out in the future. With vibe coding I was able to build a browser extension in 45 minutes that had more features than I originally planned for.

I spend more time documenting the experience than building which is wild. If you are interesting you can look at the README in https://github.com/mattheimer/vibe-headers

But I left the experience with two thoughts.

Even seasoned developers will be using vibe coding in the future.

I think in the near future the browser plugin market will partially collapse because eventually browsers will build extensions themselves on the fly using natural language.

A tree cutting tool.

Take photos of the tree from 6 different angles, feed into a 3D model generator, erode the model and generate a 3D graph representation of the tree.

The tool suggests which cuts to make and where, given a restricted fall path (e.g. constrained by a neighbors yard on one side).

I create the fallen branches in their final state along the fall plane, and create individual correction vectors mapping them back to their original state, but in an order that does not intersect other branch vectors.

The idea came to me as a particularly difficult tree needed to come down in my friends yard, and we spent hours planning it out. I've already gotten some interest from the tree-surgeon community, I just need to appify it.

Second rendition will treat the problem more as a physics one than a graph one, with some energy-minimisation methods for solving.

Robot walking. But really, a programming language and tools to make it easy to solve hard nonlinear control problems like walking. There are videos & source code for 2-legged walking robots (in simulation) at https://throbol.com/post/how-to-walk, and videos & source code for 4-legged walking robots (on real hardware) at https://throbol.com/

I’m working on Popgot (https://popgot.com), a tool that tracks unit prices (cost per ounce, sheet, pound) across Costco, Walmart, Target, and Amazon. It normalizes confusing listings (“family size”, “mega pack”, etc.) to surface the actual cheapest option for daily essentials.

On top of that, it uses a lightweight AI model to read product descriptions and filter based on things like ingredients (e.g., flagging peanut butter with BPA by checking every photograph of the plastic or avoiding palm oil by reading the nutrition facts) or brand lists (e.g., only showing WSAVA-compliant dog foods). Still reviewing results manually to catch bad extractions.

Started this to replace a spreadsheet I was keeping for bulk purchases. Slowly adding more automation like alerting on price drops or restocking when under a threshold.

https://github.com/colibri-hq/colibri/

Colibri—a self-hostable web application to manage your (and your family's) ebook library, intended as a companion to Calibre. I want it to be a friendly, simple, capable, opinionated app to review your books, add metadata to them, get them onto your reader, share them with family and (few) friends, create a public shelf for bragging, connect with Goodreads etc., and exchange comments and reviews on books.

This is explicitly not intended to ever be monetised, and I enjoy all the implications that has on the design. Colibri is as much a tool I personally want to use, as it is a study in small-audience user interfaces, and the quest to build the perfect book catalog schema.

I'm looking for fellow book-loving people to work on Colibri, to create the best personal digital library possible. If you're interested, feel free to reach out via email (in bio), or on GitHub.

Working on a web app builder that generates code via AI, much like hundreds of other tools out there. The differentiator is that the tool automatically sets up a Postgres DB (using Neon) for you. So, it's a lot easier to get started and it can handle large complex apps that require auth and database, but it can also build simple websites. The stack is next.js and code is easy to export and view.

Primarily uses Claude Sonnet 3.7 and Gemini Pro 2.5. But you can choose other models too.

You can try it for free while I'm beta testing it here: https://lumosbuilder.com?ref=hn

A few things:

- Snapshot (https://apps.shopify.com/snapshot): AI generated product photos for Shopify. Previously used Flux and Stable Diffusion but always had quality problems. Was very tricky ensuring text remained the same and the product fit into the generated background. Just integrated the new OpenAI image generation model and results are much better however their masking feature doesn't work properly so need to wait for them to fix that before I can offer the same feature of keeping text/fine details the same

- Lurk (https://apps.shopify.com/lurk): New one I just launched - allows Shopify merchants to track the prices of competitors and adjust their own in response with dynamic pricing rules. It's cool because you just have to paste a URL and it uses AI to figure out the price. Again, there's a surprising amount of things you need to figure out to make this work reliably at scale (e.g. popups, ambiguous HTML, location-based pricing, etc.)

- Origin UTM Tracking (https://apps.shopify.com/origin-utm-tracking): Simple UTM analytics for Shopify stores. Acquired this last year and its being growing nicely.

https://mysukari.com - A Diabetes management platform

I got diagnosed with type 1 diabetes in Feb (technically LADA as it's late onset). I'm the first in my family with it so I had zero info on it. I tried getting some CGMs to use but most don't work in Kenya as they are geo-locked, and even apps for measuring carbs like CalorieKing are not available in my region. I was really frustrated with the tech ecosystem, and started working on My Sukari as a platform of free tools for diabetics.

I mostly get time to work on it on the weekends, so it's not yet ready for public use, but I've fully fleshed out one of the main features: Sugar Dashboard - A dashboard that visualises your Glucose data and helps you easier analyse it.

To help with demos, I've shared my Sugar Dashboard here: https://mysukari.com/tools/sugar-dashboard/peter

I'm really passionate about this and getting as much free, practical tools in the hands of patients (it honestly shouldn't be this hard to manage a disease)

I'm working on an Ephemeral Markdown Paper: https://github.com/unvalley/ephe

If you’re tired of bloated to-do and note-taking apps, give this a try. It's OSS, free and No sign-up.

Creating a modern development environment for win98. Currently I’m working on a git client. I have no patience for learning C. So backporting git is not an option. I also don’t want to use cygwin. So I’m using a server to expose git as http endpoints and coding a git client in php to use in msdos. I have Vim 7.3 and and gnu coreutils working in msdos already. So soon I will have a very nice dev environment. I want to create a one click installer which gives you xampp, vim and then all the tooling I’ve created. I’m also interested in creating SPA that works in IE5.5. But I’ll do that when my tooling is ready.