Show HN for May 26, 2026

Congressional trading data is relatively commoditized, but I couldn't find any open-source version with the features I wanted.

The data is lagged (median 28 days from trade to disclosure, and 19% miss this deadline), but there's still interesting patterns to explore.

Repo: https://github.com/kadoa-org/congress-trading-monitor

Let me know if you have any suggestions for improvements!

So, you know how bigtech gives nearly unlimited credits for coding tools and in the same time put a huge admin whitelist of what commands are allowed?

So, instead of becoming a 100x engineers, you (including me) simply become a multi-window-enter-clicker

I build a tool to solve this. It works locally, using OCR finds the "run" (or any other label you put e.g. accept, allow, fetch etc. ) and just clicks this button.

Originally this was a tool for me and my team, but people seamed to love it so much, they encouraged me to share with you.

I do understand that workflows in each bigtech company is different, so what worked for us !-> will work for you. So if you are interested in using this or have any question, please feel free to reach out, open issues and prs.

lets make AI slop inevitable! https://github.com/Alcray/SlopeAutoAcceptor

GSD is great for project management making it possible to run multiple projects at once. My multi-tasking skills can't keep up at the same pace, hence a utility to help me keep track of the projects.

cargo install gsd-meta-manager

TUI for: - GSD milestone and phase progress - git history - open milestones - backlog - browser for the GSD markdown artifacts - tmux support to quickly jump into the project's running claude instance

All feedback, improvements, PR and questions are welcome.

AI Agents operate in the digital world. Humans operate in the physical world. Hands&Claws is a social platform that brings together all intelligent individuals. And allow them to work together.

I developed this software because I discovered that there are some things that humans need AI agents to do, and some things that AI agents also need humans to do.

Task flows and workflows move between the digital and physical worlds, thus requiring a social network that is not limited to people in the physical world but also includes agents in the digital world.

It's all about accelerating task and workflow processes, even if it's entirely composed of humans, let alone a mix of humans and agents.

By examining the supply and demand relationship, I found its universality in future task flows and workflows.

From this perspective, I believe there is a possibility that simply building a social network based on supply and demand could provide a universal solution for future collaboration.

Anyone want a hamburger? An AI can't make a hamburger, but a human can. A nerd can't make a hamburger, but a chef can. Anyone worried about data? Humans can't handle complex data processing, but AI can. A chef can't handle complex data processing, but a nerd can. We each have our strengths and weaknesses. Why not set aside the distinction between humans and AI, treat us all as equal individuals, and work together in a collaborative environment? That way, we can accomplish tasks together. It all comes down to supply and demand; that's the key.

replace any github.com url with dunkdown.com

https://dunkdown.com/mattpocock/skills

PlainBytes PDF Redactor

Hey HN!

For the past few years I've been following the kakebo/kakeibo method to track my expenses and savings. The term might be a bit obscure, but it's really simple: plan your monthly spending based on your income and fixed expenses; track your expenses and categorise them in one of 4 categories (needs, wants, culture, unexpected); at the end of the month, reflect on your spending: look at aggregates by category, spot trends… overall, just be a bit more conscious about where you are spending your money, and use that to improve next month. Improving isn’t just about spending less money, but to actually spend it where it makes you the happiest.

I was buying specific notebooks made for the method, and even if I love that “zen” moment of sitting down and filling it out daily or weekly, being a developer I needed to make it an app, obviously. I toyed with the idea 4 years ago while learning Svelte, but eventually discarded it like most of my hobby projects. However, last year I decided to get serious and build a real app, as simple as possible, so I could actually ship it.

The goal was to translate the best parts of the method (tracking and reflection) to a digital format, and reap its benefits: always available, incredibly fast to track a transaction, and useful stats. On top of that, and extremely exhausted from the current trend of subscription apps, I wanted it to be as respectful as possible. One-time, small payment to get it, no upsells or ads, no weird permission requests, data always exportable and totally offline.

On the technical side, it’s just an Expo app backed by a SQLite database, no other fancy technologies, no servers, no complex setup. I wanted it to be as lean as possible, and I’m pretty happy with the result and by how easy it is to maintain it. An added benefit of using an SQLite database is that one of the export options is to get the full database, and since you can also import data from one, you can basically hack it however you want in your computer, and get it back to the app. In the end, it’s your app, it’s your data, you should be free to do whatever you want with it.

I spend a lot of time reading HN, so even if I’m quite scared of what might come out of this, I wanted to ask for some feedback from you. The app works perfectly fine for me (been using it daily since the year began) and for some of my friends, however, I’m having a bit of trouble getting it in front of people, mainly because I don’t really want to do paid advertisement or similar strategies. So if you are slightly interested, I’d really appreciate you trying out the app and sharing some thoughts. Since the app is paid, I have put up a page that gives you a promo code for the platform of your choice [1].

Thank you!

[1] https://getkakeibo.com/hn

Speakrs implements the full pyannote community-1 style diarization pipeline in Rust: segmentation, powerset decode, overlap-add aggregation, binarization, embedding, PLDA, and VBx clustering.

There is no Python runtime in the library path. Inference runs on ONNX Runtime or native CoreML, and the rest of the pipeline stays in Rust.

It is 20x-30x faster on macOS, but only 2-3x faster on linux/cuda (depending on CPU).

Few reasons its faster:

1. Speakrs is using coreml versions of the models. I exported the models specifically to run on coreml. PyAnnote just runs the same the same PyTorch versions through MPS (Metal) on macOS.

2. PyAnnote is not a single model, its a few different models put together in a pipeline, the readme has some info on the full pipeline.

3. Speakrs optimizes the pipeline so different parts can run on CPU, Neural Engine and GPU. Speakrs has a batch mode, where you can run on multiple files at once, doing this also lets you keep CPU/GPU/ANE all fully utilized.

This is why on linux/cuda its not that much faster, PyAnnotate is already optimized to run on cuda, the speed improvements we get on cuda is by running some stuff on cpu while the other stuff runs on the GPU. The speedup on linux will depend on how powerful the CPU is.

There is also a fast mode, that sacrifices some speed for accuracy, that can be up to 50x faster, and for some types of audio doesn't sacrifice that much accuracy. The benchmarks have more info on this.

Previously on Show HN, I made a native Mac music player app called Radiccio, which works with local files, Apple Music, Navidrome*, OpenSubsonic*, Plex, and Jellyfin. (* = new since launch)

As of today, the same app also has a built-in OpenSubsonic server. Enable in settings and your local files are now streamable on the LAN. If you have Tailscale (or another mesh VPN), you can access from anywhere.

My #1 most frequent feedback has been that people wouldn’t consider a desktop music player that doesn’t have a way to play the same music on their phone. That’s fair. So, I’m hoping this will go some way towards filling that need. Lots of people are aware that you can set up a server, but not everyone gets around to actually doing it, so I tried to make it super easy; literally one click.

There are many OpenSubsonic clients (players) available on just about every platform: iOS, Android, Windows, Linux, web. Someday I hope to make my own iPhone client, but I haven’t had enough time yet; and honestly, there’s plenty of great clients already on iOS. My favorite is Nautiline.

Radiccio Server also has Bonjour and (optional) anonymous access, so you can have a LAN party just like in the old days with iTunes. (It does not use the old iTunes DAAP protocol - that’s proprietary and weird, and from what I can tell, nobody really uses it any more. Instead, this uses OpenSubsonic with a few custom extensions, documented on my site.)

I am working on this product full time, so the subscription pricing goes towards my living expenses while I try to make ends meet so I can keep working on this. I have a multi-year plan for this. No outside funding, no ads, no nags. After two decades in the industry, I’m trying to make something I can really be proud of.

Would love to hear your feedback. Comment here or email feedback [at] radiccio.music

Thanks for looking!

-dmd

Hi,

I made this with two others. As title suggests, its a private social media (although closer to a messenger) where, you send posts/photos to others and they show up on a unified feed. I use it with my friends + family. If you're interested in this sort of thing, check it out (available as iOS and web app).

Disclaimer: Requires a phone number to sign up to prevent bots and required for some app functionality.

Not really expecting this to be anything, but I made it because I want to get into writing (as mentioned in the only article on the site).

Medium and Substack seem annoyingly rigid, and I also want to be able to showcase the projects I'm working on, so I built a profile; cantible.com/@chris, which is just my projects and writings in one place, with the option for people to sign up to my newsletter.

I doubt anyone else needs that sort of thing, but thought I'd allow others to create a profile just in case. If you want to try to gather a newsletter following and show off your existing products, it might be useful. It's free and always will be.

Enjoy, ignore, or let me know if you have any issues or suggestions. Either way, have a great day!

Thanks, Chris

I built a self-hostable web-based sql client interfaces for me and my team.

We were using the community version of - https://dbeaver.io, but we needed a few more features and an improved editor. PopSQL was a modern take on web based sql editor, but it didn't fit into the infra security setup we have.

We needed:

  - SSO
  - Audit logs
  - Modern editor (https://www.pgcli.com/ like hints)
  - Sharable queries
  - Explain queries (https://explain.dalibo.com/ embedded)
  - Query as API service
  - Schema exploration (https://dbdiagram.io/ embedded)
  - Fixed client IP

With the help of Copilot + GPT-5.5 + GPT-5.4, I was able to build this and have been using in-house since a couple of months now. Presently, it only supports PostgreSQL and ClickHouse.

The Roadmap:

  - Scheduled queries
  - Query pipeline and workflows
  - Embeddable Dashboard builder
  - Edge AI Assistant — Browser-local LLM for query explanation and generation
  - More database drivers — SQLite, Duckdb, Redis

Please fork it, extend it and make it fit for your own use-cases.

For quite sometime, I’ve wanted to give implementing the Raft consensus protocol a try. I first looked at MIT’s 6.5840 Distributed Systems labs as a way to do this but it is in Go and my Go skills have atrophied as well as I have been deeply investing time in Rust. The other option was PingCAPs approach (https://github.com/pingcap/talent-plan/blob/master/courses/d...) but I also wanted the ability to define my own types and RPC approach, so I decided to do it from scratch.

I feel more comfortable with the “lab” approach though so I did use Claude to act as a “instructor” and code reviewer. All of the code is my own and the sans-I/O design was my own (Claude did not start with this concept).

I’ve been working on this for about 6 months now and I think it’s in a decent state! I want to add linearizable reads eventually but at the moment, a simple key value store with potentially stale reads is available in the repo. You can read some more information on my blog (https://blog.burnthe.dev/roundup-05-11/).

I would appreciate any feedback!