Ask HN: What Are You Working On? (June 2025)

I'm working on a new app for creating technical diagrams - https://vexlio.com. It's an area with some heavyweight incumbents (e.g. Visio, Lucid) but I think there's good opportunity here to differentiate in simplicity and overall experience. I'm still in the fairly early phase, and I suspect I haven't quite found the best match of features to customers yet.

From a dev perspective this area has a ton of super interesting algorithmic / math / data structure applications, and computational geometry has always been special to me. It's a lot of fun to work on.

If anyone here is interested in this as a user, I'd love for any feedback or comments, here or you can email me directly: tyler@vexlio.com.

Some pages the HN crowd might be interested in:

* https://vexlio.com/blog/making-diagrams-with-syntax-highligh... * https://vexlio.com/solutions/state-diagram-maker/ * https://vexlio.com/blog/speed-up-your-overleaf-workflow-fast...

I _was_ working on an open-source, self-hostable app for sending out newsletter to your friends and families. I made a MVP but then I scrapped it after realising how cumbersome it is to manage email related functionalities. Since its strictly for connecting with your friends and family, I figured, why not let users use their own email to send out the updates.

So I made a proof of concept app on iOS that uses gmail API to send out newsletter emails. I wish I could just send prepopulated emails (with inline attachments and recipients) to iOS mail client instead of asking for gmail OAuth permissions, but it doesn't look possible.

Now I'm trying to create a polished app for alpha testing. Been exploring data persistence (Swift Data, Core Data, rxdb etc) and settled on Core Data. Architecture wise, I've settled on MVVM + Swift UI. At the moment I'm trying to figure out how to make mocks and XCode preview data geneeration ergonomic.

So far, I am pleasantly surprised at Swift and iOS development, but I still hate XCode.

A homegrown Plex.

After a lot of grief trying to make Plex and jellyfish to work with my collection, and then some more with the community [1] I decided to make my own.

There's no selling point and clear pathway to monetize, as other solutions are way more mature and feature complete, but this is my own and serves my needs the best.

I've been working on it on and off for last 8 years or so, and it's been my personal benchmark for js ecosystem. The way it works, every now and then I come back to the project, look at the latest trends in js world and ask myself a simple question - what should I change in the codebase to make it online with the latest trends. And everytime it leads to full rewrite. Kind of funny, kind of sad.

In a nutshell I have a huge movie collection - basically I'm preparing for armageddon where all online streaming services cease to exist and I need both backend to fetch me detailed information about movies in the collection as well as frontend to help to decide what to watch tonight.

My next major endeavor will be trying to integrate RAG to take a bite at my holy grail - being able to ask a question like "get me a good gangster flick" and get reasonable recommendations.

[1] I think it was jellyfish where I was asking on their forums for how to manually create a collection, stating I'm a software engineer with 20+ exp and they kept telling me that I shouldn't touch the code... While having an online campaign asking for volunteers to contribute to the codebase.

I’m working on a native swift implementation of FileKitty, the FOSS LLM prompt context preparation tool I’ve been building with pyqt.

https://github.com/banagale/FileKitty

My most recent release includes signed .dmg installer on top of brew, and a local build option.

Although it should compile to any platform, I want to take advantage of the new Foundation Model sdk Apple announced at WWDC.

I also recently released something called slackprep, a CLI tool and Python library that wraps slackdump, converting Slack export data into LLM-groomed Markdown transcripts.

That includes labeling inline images organizing them for upload as LLM context.

https://github.com/banagale/slackprep

I see these and other utilities coming together to assist in assembly of deep context for system level design.

I'm working on a system to help people write their Family History.

You upload interviews with family members (text, audio or video all work) and the system automatically transcribes the text, finds key people or events, and puts it together with other information you may have gathered about those events or people before. Like building a genealogical tree but with the actual details about people's lives.

In the works to also attach pictures of said people and events to give it some life.