Show HN for July 19, 2025

Hey HN,

I am Cam, and for the past two years I have been working on Compass, an open-source handheld CNC router that brings computer precision to woodworking while keeping the user directly involved in the process.

The idea started as my senior design project at UC Berkeley, with the goal of making a more approachable CNC machine—standard CNC machines are expensive, bulky, and remove you from the tactile “maker” experience. Compass solves that by combining a handheld router with real-time robotic assistance. You move the router roughly along a design path, and Compass uses four optical flow sensors (like in computer mice) and a 3-axis motion system to auto-correct for precision cuts.

What is different about Compass: - Open source: All plans, firmware, and CAD files are available on GitHub. - Affordable: The DIY build costs ~$600 in parts, and I am selling kits for <$800. - No external markers: The sensing technology allows for positioning without external markers, so no setup or consumables required. - Portable: Fits in a backpack and is not limited by a fixed work envelope.

We recently completed our first beta program and have just launched V1 kits for pre-order. You can find more info and the launch video at the listed URL.

GitHub: https://github.com/camchaney/handheld-cnc

I think it's finally time to lock version 0.2 of my 4d side project. (Since previous Show HN a couple of years ago: https://news.ycombinator.com/item?id=37100760 I scrapped lots of stuff and concentrated on making a javascript "game engine")

The sole purpose of this project is to explore whether it's possible to make the human brain spatially aware of four dimensions using Virtual Reality. It’s my first ever game project, my first time using Unity, first time writing C#, and my first anything with WebXR, so please be gentle! To manage expectations, it isn't visually pleasing yet and it's not easily approachable. To get anything out of this, you likely need a pen and a paper and draw 3d slices of 4d objects _many_ times.

The project is essentially a rudimentary engine where you can use JavaScript to define 4D objects and some limited behaviors. You can then explore these spaces in a browser (with or without a VR headset), moving and rotating through all four dimensions—including double rotations on planes like xz and yw.The game is a webxr project, meaning you can just play that on the browser, even without VR headset. Note, however, that the keyboard input is a bit limited.

As a demo, I've built a simple maze game where you try to find gems before time runs out. The challenge ramps up, and I'm currently stuck trying to comfortably navigate Level 6, which involves moving around four separate tesseracts. Here is a page where I have tried to visualize what is going on at level 6: https://www.brainpaingames.com/blog/updates/2024/07/28/Hyper...

I have no expectations that this will any day be anything but a niche project of mine, but I thought here might be someone interested in playing around. And If anyone has any ideas how the pedagogy of learning to operate in 4 spatial dimensions could be made somehow more achevable, I'm all ears.

A while back, I built a simple app to track stocks. It pulled market data and generated daily reports based on my risk tolerance. Basically a personal investment assistant. It worked well enough that I kept going.

Now, the same framework helps me with real estate: comparing neighborhoods, checking flood risk, weather patterns, school zones, old vs. new builds, etc. It’s a messy, multi-variable decision—which turns out to be a great use case for AI agents.

Instead of ChatGPT or Grok 4, I use mcp-agent, which lets me build a persistent, multi-agent system that pulls live data, remembers my preferences, and improves over time.

Key pieces: • Orchestrator: picks the right agent or tool for the job • EvaluatorOptimizer: rates and refines the results until they’re high quality • Elicitation: adds a human-in-the-loop when needed • MCP server: exposes everything via API so I can use it in Streamlit, CLI, or anywhere • Memory: stores preferences and outcomes for personalization

It’s modular, model-agnostic (works with GPT-4 or local models via Ollama), and shareable.

Let me know what you all think!

I launched an iOS app that lets you scan documents as quickly as your camera takes photos, all wrapped in the skeuomorphic style of iOS 6.0.

It was a fun indie project. Hope some of you will enjoy it! :)

Hi HN, the last half year I spent some of my spare time on learning Vulkan (mainly the compute part) and applied it directly in developing a gaussian splatting renderer. I am not sure where I will go with this, but I thought some of you might find it interesting and/or have suggestions on how to proceed.

Ultimately, I want to Klartraum to be an extensive neural network and rendering inference engine (no training, backprop, autograd) running on embedded devices and VR headsets with high performance. My major obstacle right now is that writing GLSL compute kernels is tedious compared to CUDA ...

What do you think should be added so the library would be of use for others?

Hey everyone!

I've been a Splitwise user for a long time (sometimes paying, sometimes not), and have been rather frustrated at their new-ish policy of only allowing N (small) expenses to be submitted per day. And I was curious to write some Rust and thought the expense simplification algorithm would be fun to implement, so I built this.

Thought others might find it fun or useful! It's dead simple to self host if you want to try it out :)

I've been self-learning from YouTube for years—everything from coding to design to business skills. But I kept hitting the same wall: YouTube learning has no structure. Your knowledge gets scattered across random playlists, you're passively consuming content without real retention, and when you're confused, there's nobody to ask.

I built Notetube to fix this by layering organizational tools with AI to create a proper learning system:

Organizational layer: Build structured collections by topic/course/skill, visualize your learning progress with dashboards, create and track your learning goals

AI layer: Automatically generates detailed notes (3000+ words for 1 hour of content) and summaries, identifies key moments with timestamps, creates personalized quizzes for retention testing, and provides a chat interface for instant help when concepts aren't clear

...plus additional features like timestamped note-taking, but I'll keep this brief.

Quick signup via Google OAuth for a smooth onboarding experience.

Try it free: https://www.notetubeai.com/

Would love your thoughts and feedback from the HN community!

built a simple study app that turns messy notes (images or text) into flashcards + quizzes using gpt-4o.

originally made it to help myself focus while studying. launched with a hard paywall to avoid ads or noise.

~60 users so far, mostly from tiktok. curious if this is useful beyond my own use case?

A Chrome extension that provides Vim-like keyboard navigation for search result pages.

Pressing Enter immediately after searching opens the first search result instantly.

j key: Select next search result k key: Select previous search result Enter key: Open selected search result Highlight selected search result Auto-scroll functionality