Hand - Moozonian Search

github.com• Jul 14, 2026• 1 min read

Show HN: GLP-RAM – Chrome extension reducing the browser memory appetiteHi!I've been struggling to use Chrome on my M1 Mac, because of how much RAM it eats, and memory saver wasn't enough.I've made this extension based on combination of features on multiple separate extensions: 1. When you click on the new tab, only the first tab loads in background, the rest will have a placeholder until you click. 2. Only 3 active tabs per window are allowed to be loaded, the rest will be killed. 3. The extension watches if you have unsaved input, paused or active media, or if the page is configured to send notifications and it keeps these tabs alive.I've been using it for a couple of weeks and it allows me to open way too many tabs to handle :)The extension is published in Chrome Web Store, let me know if you have any questions!

news.ycombinator.com• Jul 14, 2026• 1 min read

Show HN: I can't seem to text properly on an iPhone, so I built my own keyboard== The problem ==I’ve never been able to text properly on a phone. I’ve tried with two thumbs, one thumb, and every other way I can think of. Nothing works for me. I can only seem to text with my right index finger while holding the phone in my left hand, which is at best is slow and full of typos. I find myself fighting autocorrect as much as using it. I tried swipe keyboards (gboard, etc.) but they didn't help. Sometimes I watch people texting rapidly with both thumbs and marvel at their dexterity. It's all very frustrating for me.== A solution ==So I set out to build a custom iPhone keyboard designed around one idea: making the keys as large as possible.Initially, I wanted to build a transparent keyboard that would overlay the whole screen, but iOS doesn't allow this. So I made as much room for larger keys in the allowed keyboard real estate by:* Removing frequently used keys (spacebar, backspace and shift), and replacing them with gestures* Removing keys for accessing focused keybo

bing.com• Jul 14, 2026• 1 min read

Mom of boy who froze to death after dad forced him to sleep in the garage agrees to settlement instead of a civil trialThe mother of a New York boy who died at the hands of his father and his fiancee has indicated that she would accept a settlement over a civil trial. Justyna Zubko-Valva, the mother of 8-year-old ...

stagewhisper.io• Jul 14, 2026• 1 min read

Show HN: BYO AI free notetaking with optional screen reading for OpenClaw/hermesI've built fully on-device (macOS) meeting transcription/summaries/action-items app. It uses parakeet and gemma 4 or OpenClaw or Hermes Agent to drive main functionalities. It's free, requires no account or email, and is source available. You can also natively connect your own local/VPS OpenClaw or Hermes to use them, their memory and context, as a notetaker.About me: my name is Piotr. I'm having plenty of calls as a freelancing programmer, and have interest in audio domain, so it was natural for me to pick this as my side project. I struggled a bit with echo cancellation, and writing idiomatic plugins to let users use their OpenClaw or Hermes agents as BYO AI (basically backend). I started project coding Tauri app by hand, only after establishing architecture I started leaning on AI coding agents. I use Claude Code with https://github.com/openai/codex-plugin-cc for "adversarial-reviews".On the website I've included an offer to upgrade StageWhisper Lite to Founders edition with one-tim

news.ycombinator.com• Jul 14, 2026• 1 min read

Why can't you commit a .env file?"You can't commit a .env file to Git" is a sentence we've all heard. Most people accepted it and went on with their lives, using other tools to share environment variables.But the question we should ask is: why can't I commit a .env file to Git?The honest answer isn't that it's bad practice. It's that Git has no permission system below the repository level. It's all or nothing. You see everything, or you see nothing. The entire secret-manager industry exists to paper over this one missing primitive.For too long we accepted it as a minor annoyance. That's changed. It isn't a minor annoyance anymore. It's a real problem.We now have agents monitoring every patch that merges, hunting for security fixes to turn into exploits. The patch itself is the disclosure: it hands people (and increasingly agents) everything they need to reverse-engineer the fix and hit systems that haven't updated yet. We're in the middle of a security crisis, arguing about where to store files to hide them from attac

news.ycombinator.com• Jul 13, 2026• 1 min read

Web Client SDKThe Atoa Web Client SDK lets you embed a seamless payment experience into your web checkout in minutes. It provides a simple interface for a payment dialog supporting both pay-by-bank and card payments (Visa, Mastercard, Apple Pay and Google Pay), handling the entire flow securely and efficiently — whether you’re building a custom payment gateway or enhancing an existing one.

github.com• Jul 13, 2026• 1 min read

Show HN: AffineUI a 2 file C++ GPU/HTML5 UI that can replace ElectronI thought this might be useful to C++ engineers working on games, graphics, content in C++ looking for a good lightweight html5 Electron style UI engine. AffineUI is deeply optimized and meant to run as close to 120hz as possible, but supports (most) of html5 and is easy to integrate with your app. It also is highly crash resistant using weakrefs, handles, etc. to avoid footguns.A zero-dep drop in lib that could be a cross between DearIMGUI and Electron. A full featured GPU accelerated html5 rendering engine you can drop into any C++ project with two files (cpp/h) that adds a gpu accellerated html5 Electron style render plus a Gradio style UI component framework.Just main.cpp, affine.cpp and affine.h and you have yourself an app that's about 1/100th the size of the average Electron beast. I built it as a nicer (but somewhat larger) replacement for DearIMGUI that had full CSS/HTML5 compatibility for graphics apps I've been working on.Also available as a python library (pip install affin

bing.com• Jul 13, 2026• 1 min read

Satya Nadella has issued a shocking warning to companies using AIOf all the debates raging about the potential downsides of AI, there is one worry causing the most hand-wringing among AI enthusiasts in Silicon Valley — that the giant AI labs that sell proprietary ...

news.ycombinator.com• Jul 13, 2026• 1 min read

Show HN: TokenDrift – find hardcoded colors/spacing hiding in your CSSHi HN — I built TokenDrift, a CLI that scans a codebase for design system drift: hardcoded colors and off-scale spacing that quietly bypass your design tokens.npx tokendrift .It's zero-dependency Node, runs entirely locally, and gives you a Drift Score (0-100) plus a report showing exactly which files and lines are responsible. --share is opt-in and uploads only the report data (never source), if you want a link to paste in a PR or Slack.Why I built it: every design system I've worked with rots the same way. Someone hardcodes #3B82F6 under deadline pressure, someone else does margin: 13px instead of the actual spacing scale, and six months later nobody can say when "mostly following the system" became "barely following it" — because nothing was measuring it.Three real examples, not staged demos — I ran it against open-source repos and checked every number by hand before publishing: - Dub (dub.co): 82/100, 93% token adoption https://tokendrift-vedantyedes-projects.vercel.app/r/9GNaJ2M..

news.ycombinator.com• Jul 12, 2026• 1 min read

Ask HN: Would filesystem bookmarks be useful in your shell workflow?I’m experimenting with a small shell utility for filesystem bookmarks, and I’d like to understand whether this is a problem other people have.The idea is to define persistent or temporary bookmarks for directories:``` bm add work ~/Work bm temp build /tmp/build ```and then use them directly in paths:``` cd @work ls @work/project ```The utility would also support renaming bookmarks, promoting temporary bookmarks to persistent ones, demoting them, and clearing temporary bookmarks.I’m not sharing the implementation yet. At this stage, I’m mainly interested in how other people handle this today:* Do you use aliases, environment variables, shell functions, z, autojump, or something else? * Would explicit names such as @work be useful, or would they feel redundant? * Would temporary bookmarks be useful in real workflows? * What would this need to offer to be worth using instead of existing directory-jumping tools?

news.ycombinator.com• Jul 12, 2026• 1 min read

Ask HN: Bypass smartphone call apps? Silent VoLTE via RTP extensionhttps://github.com/sigureitimonnji/nvap-protocol [EN] I designed a protocol called NVAP to eliminate the need for vocalizing personal information during call center verification. Check docs/sequence.md in the repo for the core timeline logic.To bypass the strict API/kernel-level RTP restrictions of iOS/Android, I'm thinking of completely bypassing the native call app and handling it via the earphone layer (utilizing RFC 8285 / 9335 Cryptex).However, I am facing two practical challenges regarding the timeline and biometrics:1. The Inbound/Outbound Timeline Mismatch: If the other party answers instantly on an outbound call, or if the user presses the answer button before putting the phone to their ear on an inbound call, the VoLTE connection starts BEFORE the earphone is actually in contact with the ear, ruining the acoustic scanning timeline.2. Biometric Ownership: If the master biometric data is inside the smartphone's Secure Enclave, transferring data back and forth over Bluetooth int

github.com• Jul 12, 2026• 1 min read

Show HN: Itara – Distributed system topology as an explicit, executable layerHi HN, I'm Gábor, a software engineer from Budapest.I spent almost a decade designing, building and maintaining distributed systems, and I came to truly understand why a lot of people define software architecture as "the stuff that's hard to change later". Changing service boundaries, communication protocol or serializers is time consuming and risky. The past few months, I've been building Itara, my attempt to ease that pain.Itara takes the software topology that's currently spread across the codebase and configuration and infrastructure, and concentrates it into a dedicated, executable layer. This dedicated layer describes the topology as a directed graph, where the nodes are the components of the system, and the edges are the connections between them. The edges have all the properties of the connection, like the transport to use, the serializer to use, the failure handling strategy and so on. Colocated components, components running in the same process, are modelled with direct conne

kurvengefahr.org• Jul 12, 2026• 1 min read

Show HN: Kurvengefahr – browser CAD/CAM for pen plottersA few years ago I made a pen plotter attachment for Prusa MK4 (https://www.printables.com/model/827264-pen-plotter-attachme...) and at the time I didn't have a good way to turn artwork into G-code for it, and I put the project on ice for a while.I recently wanted to dabble in line art again and made a small browser app to make it easier. As agentic AI tools of 2026 are quite addictive, it rather quickly grew into something quite a bit more - an integrated browser CAD/CAM for pen plotters that covers everything from importing existing artwork, creating artwork from scratch, preparing for plotting and hardware integration. It includes some off-beat features like a Logo interpreter for turtle art and Graves RNN for handwriting synthesis and in addition to 3D printer pretending to be pen plotters it now also supports actual pen plotters based on EBB (AxiDraw) and GRBL firmwares through Web Serial.If you own an AxiDraw or a GRBL plotter, I'd very much appreciate it you gave it a try and giv

github.com• Jul 12, 2026• 1 min read

Show HN: Skillscript – A declarative, sandboxed language for tool orchestrationHi HN — I'm Scott. Skillscript is a small language I built to write what I want my local agent to actually do, in a form I can read and version, instead of hoping the model gets it right each time.The itch started with something small. I wanted my NanoClaw agent to run my morning brief the same way every day. Check overnight tickets, summarize the deploy pipeline, flag anything urgent. Every session, it would re-figure out how to do this from scratch, drift a little, and cost tokens for what's basically a fixed procedure. I could put it in a system prompt or an MD skill file, but those are still instructions the model reads and reasons about every time. And I wanted it to run autonomously and then hand it to the model to reason over the data.The second thing that pushed me: I wanted to use small local models for the cheap stuff. They're capable, but if you just hand them the wheel, they wander. What I wanted was a way for the frontier model (or me) to write a specific procedure and han

github.com• Jul 12, 2026• 1 min read

Show HN: Levee – a self-tuning circuit breaker and concurrency limiter for GoRate limiters and circuit breakers work wonderfully well when they’re configured for the load and available capacity. But load, capacity, and latency drift over time, and keeping those settings current requires continuous effort. So I built Levee to be a hands-off, adaptive, easy to configure traffic governor.Levee is configured with a success-rate target and timeout. It then continuously monitors the workload performance characteristics to detect downstream capacity exhaustion or failures. It can also spot a surge from growing concurrency before failures arrive. It runs in-process, uses a small, fixed amount of memory, has zero dependencies, and processes millions of requests per second.In a deterministic 10-node mesh simulation, Levee outscored carefully tuned static rate limiters/breakers deployed throughout the mesh, while recording fewer failures and no node crashes.

github.com• Jul 11, 2026• 1 min read

Show HN: A deterministic I Ching engine, cross-validated against another implI built a deterministic engine for I Ching (六爻 / Liu Yao) hexagram casting and open-sourced the core (Apache-2.0). I want to be upfront that the interesting part here is not the divination — it's a testing problem, and I'd like feedback on how I handled it.The engine takes a moment in time plus a coin toss and mechanically derives a fully annotated chart: the hexagram, its palace, the stem-branch (najia) assignments, the "six relatives," moving lines, void days, and the line-strength skeleton. It does not interpret anything — it only emits hard, mechanical facts.The problem that made this interesting: the rules are ~3000 years old, written in classical Chinese, and different schools contradict each other. There is no official answer key, and human experts disagree and are sometimes wrong. So the tests I'd naturally write just encode the same assumptions my code does — if I misread a rule, the code and its test are wrong together and both stay green. The self-validation paradox.My way o

news.ycombinator.com• Jul 11, 2026• 1 min read

One Wikipedia page costs your AI agent 68,000 tokensi use claude code daily and measured what pages cost it while doing research. an average wikipedia article, for instance, is 68,240 tokens of raw html (tiktoken); nike's homepage is 353,000.claude code's built-in webfetch handles the easy case well. it summarizes wikipedia to about 950 tokens and clears cloudflare on some sites like indeed and ticketmaster. but, and there's always a but, on js-rendered and some anti-bot pages it returns nothing.quotes.toscrape.com/js gives "no quotes found"; nike.com gives a 403. your agent then dumps the raw html back into context and still fails. (note: i have also had cases where i read through the chat at the end and saw that it failed and just pulled from either training data or stale caches from other sources)so i worked on building an open-source stealth browser (recompiled chromium) that runs as an mcp for claude code, cursor, and claude desktop. essentially all i have to change form my end is add the mcp, and it returns the cleaned up tokens w

news.ycombinator.com• Jul 10, 2026• 1 min read

Show HN: Cactus v2 – On-device AI with cloud fallbackHi HN, Roman and Henry here from Cactus (https://github.com/cactus-compute/cactus).We just shipped the biggest upgrade to our on-device inference platform:- Built-in model confidence-based routing to hand off inference runs to the cloud - Converter for any PyTorch model - Lossless 4-bit quantization (evals on our GitHub README) - GPU acceleration on compatible devices (starting with Apple Metal) - Minimal RAM footprint - Runs on any Arm device: iOS, Android, Mac, DGX Spark, Raspberry Pi, and moreAll in, a Gemma 4 E2B class model runs at 169 tok/sec on M5 Max, takes 2.7GB disk space with no accuracy degradation from FP16, uses 1.3GB of RAM, and requests help from cloud models when needed.The problem we started with eighteen months ago: inference engines are built for datacenters, but consumer hardware has different physics: you share RAM with the OS, you get thermally throttled, and the same model behaves differently on different hardware.So we wrote a runtime from scratch for resource-

sxp.studio• Jul 10, 2026• 1 min read

Show HN: SubjectiveZero, an open-source agentic node editor for creative codingHey there,My name is Clem, I've been a solo indie dev for a couple years now, exploring frontier tech like XR and agentic workflows in the context of creative / interactive work.I've been building creation tools for a while and some common design challenge is to figure out the right level of abstraction for your tool. You can always make it super advanced and complex with low level concepts (shader composition, actual code etc.) but then you get something with a high complexity / learning curve. On the other hand, if you make your tool too high level, it might be easier to use at first, but people will most likely hit a wall eventually and start fighting with your tool to get their edge case done (you see that on mobile a lot actually).With this prototype (called SubjectiveZero), I'd like to imagine that we can kind of move the "slider" on the abstraction layer, meaning that you can actually start with prompts that describe the goal, and you can go as high level (stay with abstract pro