🛠️ Hacker News Tools

[Other] Preventing Kubernetes from Pulling the Pause Image from the Internet

[Other] Analyzing the Performance of WebAssembly vs. Native Code

[Other] Launch HN: Plexe (YC X25) – Build production-grade ML models from prompts Hey HN! We&#x27;re Vaibhav and Marcello, founders of Plexe (<a href="https:&#x2F;&#x2F;www.plexe.ai">https:&#x2F;&#x2F;www.plexe.ai</a>). We create production-ready ML models from natural language descriptions. Tell Plexe what ML problem you want to solve, point it at your data, and it handles the entire pipeline from feature engineering to deployment.<p>Here’s a walkthrough: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=TbOfx6UPuX4" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=TbOfx6UPuX4</a>.<p>ML teams waste too much time on generic heavy lifting. Every project follows the same pattern: 20% understanding objectives, 60% wrangling data and engineering features, 20% experimenting with models. Most of this is formulaic but burns months of engineering time. Throwing LLMs at it isn&#x27;t the answer as that just trades engineering time for compute costs and worse accuracy. Plexe automates this repetitive 80%, so your team can work faster on what actually has value.<p>You describe your problem in plain English (&quot;fraud detection model for transactions&quot; or &quot;product embedding model for search&quot;), connect your data (Postgres, Snowflake, S3, direct upload, etc), and then Plexe: - Analyzes data and engineers features automatically - Runs experiments across multiple architectures (logistic regression to neural nets) - Generates comprehensive evaluation reports with error analysis, robustness testing, and prioritized recommendations to provide actionable guidance - Deploys the best model with monitoring and automatic retraining<p>We did a Show HN for our open-source library five months ago (<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=43906346">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=43906346</a>). Since then, we&#x27;ve launched our commercial platform with interactive refinement, production-grade model evaluations, retraining pipeline, data connectors, analytics dashboards, and deployment for online and batch inference.<p>We use a multi-agent architecture where specialized agents handle different pipeline stages. Each agent focuses on its domain: data analysis, feature engineering, model selection, deployment, and so on. The platform tracks all experiments and generates exportable Python code.<p>Our open-source core (<a href="https:&#x2F;&#x2F;github.com&#x2F;plexe-ai&#x2F;plexe" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;plexe-ai&#x2F;plexe</a>, Apache 2.0) remains free for local development. For the paid product, our pricing is usage-based, with a minimum top up of $10. Enterprises can self-host the entire platform. You can sign up on <a href="https:&#x2F;&#x2F;console.plexe.ai">https:&#x2F;&#x2F;console.plexe.ai</a>. Use promo code `LAUNCHDAY20` to get $20 to try out the platform.<p>We’d love to hear your thoughts on the problem and feedback on the platform!

[Other] How devtools map minified JS code back to your TypeScript source code

[Other] Show HN: Agor → Figma for AI Coding (Open Source)

[Other] Debugging containers that have no shell

[Other] Show HN: PingStalker – A a macOS tool for network engineers Hi HN - I’m the developer of PingStalker, a macOS utility I built to see what’s really happening on the LAN&#x2F;WLAN.<p>I live in the CLI, but when it came to discovery and monitoring, I found it limiting. So I built a GUI that brings my favorite tools together in one place.<p>PingStalker started because I wanted to know if something on the network was scanning my machine. I also wanted quick access to core details—external IP, Wi-Fi data, and local topology. Then I wanted more: fast, reliable scans using ARP tables and ICMP.<p>As a Wi-Fi engineer, I couldn’t stop there. I kept adding ways to surface what’s actually going on behind the scenes.<p>A few highlights:<p>- Performs ARP, ICMP, mDNS, and DNS scans to discover every device on your subnet, showing IP, MAC, vendor, and open ports.<p>- Continuously monitors selected hosts (“live ping”) to visualize latency spikes, missed pings, and reconnects.<p>- Detects VLANs on trunk or hybrid ports, exposing when your Mac is sitting on a tagged interface.<p>- Captures just the important live traffic — DHCP events, ARP broadcasts, 802.1X authentication, LLDP&#x2F;CDP neighbor data, ICMP packets, and off-subnet chatter — to give you a real-time pulse of your network.<p>- Decodes mDNS traffic into human-readable form (that one took months of deep dives, but the output is finally clear and useful).<p>- Built my own custom vendor-logo database: I wrote a tool that links MAC OUIs with their companies, fetches each vendor’s favicon, and stores them locally so scan results feel alive and recognizable.<p>Under the hood it’s written in Swift. It uses low-level BSD sockets for ping and ARP, plus Apple’s Network framework for interface enumeration. The rest relies on familiar command-line tools. It’s fast.<p>I’d love feedback from anyone who builds or uses network diagnostic tools:<p>- Does this fill a gap you’ve run into on macOS?<p>- Any ideas for improving scan speed or how traffic events are visualized?<p>- What else would you like to see?<p>Details and screenshots: https:&#x2F;&#x2F;pingstalker.com<p>Happy to answer any technical questions about the implementation, Swift APIs, or macOS permission model.

[Other] The Paranoid Guide to Running Copilot CLI in a Secure Docker Sandbox

[Other] Show HN: Sparktype – a CMS and SSG that runs entirely in the browser Hi HN,<p>After trying to teach a non-technical friend how to manage a Jekyll site I decided there must be a way to make building a site with a SSG easier. Options like Decap, Contentful etc. do make it a bit easier but still take lots of tech knowledge to set up.<p>So I built Sparktype, a browser-based CMS that outputs statically-generated HTML and CSS. My goal is for it to be as easy to use as Substack or Medium, while providing all the benefits of a static site generator including openness, simplicity, speed, security and ownership.<p>It handles most things that you&#x27;d need from a CMS, including creating pages, image resizing, menu management, tags, collections, listings etc. I&#x27;ve only made two themes so far, but I&#x27;m working on a theme store and the ability to import custom themes.<p>Content is saved as plain Markdown + YAML frontmatter and JSON config files, so there&#x27;s no lock-in and content is easily portable to other platforms. Generated sites can be exported as a zip file to upload via FTP, committed to Github or published via Netlify API.<p>I&#x27;m working on cross-platform client apps using Tauri which will enable more publishing options as its not limited by what can be done in a client-only environment.<p>The way the system works means that the Web doesn&#x27;t need to be the only interface to the content - here&#x27;s a simple Go-based CLI client that bypasses the HTML altogether <a href="https:&#x2F;&#x2F;github.com&#x2F;sparktype-project&#x2F;sparktype&#x2F;tree&#x2F;main&#x2F;st-cli" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;sparktype-project&#x2F;sparktype&#x2F;tree&#x2F;main&#x2F;st-...</a><p>It&#x27;s very early days and there are still plenty of bugs, but I&#x27;m posting now to hopefully get feedback and see what people think. Please do let me know!

[Other] OpenTelemetry: Escape Hatch from the Observability Cartel

[Other] Building blobd: single-machine object store with sub-ms reads and 15 GB/s upload

[Other] Show HN: MyTimers.app offline-first PWA with no build step and zero dependencies Hello,<p>For quite some time, I&#x27;ve been unsatisfied with the built-in timers on both Android and iOS; especially for workouts, when I needed to set up a configurable number of series with rest periods in between. That&#x27;s when I started thinking about building something myself. It was just a timer and I said to myself &quot;how hard could it be?&quot;, I had no idea.<p>The first iteration of the project worked &quot;just fine&quot;, but the UI was an eyesore (even more than it is now), and the UX was quite awful as well. As you can probably guess, I&#x27;m not versed in design or front-end development. In fact, my last real experience with front-end work was back when jQuery was still a thing.<p>However, I knew what I wanted to build, and over the last few days (and with the help of the infamous AI) I was able to wrap up the project for my needs. It required quite a lot of &quot;hand holding&quot; and &quot;back and forth&quot;, but it helped me smooth out the rough edges and provided great suggestions about the latest ES6 features.<p>The project is, as the title states, an offline-first PWA with zero dependencies; no build step, no cookies, no links, no analytics, nothing other than timers. It uses `Web Components` (a really nice feature, in my opinion, though I still don&#x27;t get why we can&#x27;t easily inherit styles from the global scope) and `localStorage` to save timers between uses.<p>I&#x27;d appreciate any comments or suggestions, since I just want to keep learning new things.<p><a href="https:&#x2F;&#x2F;mytimers.app&#x2F;" rel="nofollow">https:&#x2F;&#x2F;mytimers.app&#x2F;</a>

[Other] Show HN: AgentML – SCXML for Deterministic AI Agents (MIT) Hey HN,<p>We’ve been experimenting with how to make AI agents more deterministic, observable, and production-safe, and that led us to build AgentML — an open-source language for defining agent behavior as state machines, not prompt chains.<p>My co-founder posted before but linked to the project website instead of the repo, so resharing here.<p>AgentML lets you describe your agent’s reasoning and actions as a finite-state model (think SCXML for agents). Each state, transition, and tool call is explicit and machine-verifiable.<p>That means you can:<p>- Reproduce any decision path deterministically<p>- Trace reasoning and tool calls for debugging or compliance<p>- Guarantee agents only take valid actions (e.g. “never send a payment before verification”)<p>- Run locally, in the cloud, or within MCP-based frameworks<p>Example:<p>```<p>&lt;?xml version=&quot;1.0&quot; encoding=&quot;UTF-8&quot;?&gt;<p>&lt;agentml xmlns=&quot;github.com&#x2F;agentflare-ai&#x2F;agentml&quot; xmlns:openai=&quot;github.com&#x2F;agentflare-ai&#x2F;agentml-go&#x2F;openai&quot; version=&quot;1.0&quot; datamodel=&quot;ecmascript&quot; name=&quot;researcher&quot;&gt;<p>&lt;datamodel&gt;<p><pre><code> &lt;data id=&quot;papers&quot; expr=&quot;[]&quot; schema=&#x27;{&quot;type&quot;:&quot;array&quot;,&quot;description&quot;:&quot;Fetched papers from Hugging Face&quot;}&#x27; &#x2F;&gt; &lt;data id=&quot;summary&quot; expr=&quot;&#x27;&#x27;&quot; schema=&#x27;{&quot;type&quot;:&quot;string&quot;,&quot;description&quot;:&quot;Summary of the papers&quot;}&#x27; &#x2F;&gt; </code></pre> &lt;&#x2F;datamodel&gt;<p>&lt;state id=&quot;start&quot;&gt;<p><pre><code> &lt;onentry&gt; &lt;log label=&quot;Researcher: &quot; expr=&quot;`Fetching papers from Hugging Face and summarizing with OpenAI\n`&quot; &#x2F;&gt; &lt;openai:generate model=&quot;gpt-4o&quot; location=&quot;summary&quot; stream=&quot;false&quot;&gt; &lt;openai:prompt&gt;Summarize these recent AI&#x2F;ML papers from Hugging Face: {{fetch &quot;https:&#x2F;&#x2F;huggingface.co&#x2F;api&#x2F;daily_papers&quot;}} Provide a concise summary of the key trends, breakthroughs, and developments in AI&#x2F;ML research. &lt;&#x2F;openai:prompt&gt; &lt;&#x2F;openai:generate&gt; &lt;&#x2F;onentry&gt; &lt;transition target=&quot;log_summary&quot; &#x2F;&gt; </code></pre> &lt;&#x2F;state&gt;<p>&lt;state id=&quot;log_summary&quot;&gt;<p><pre><code> &lt;onentry&gt; &lt;log label=&quot;Researcher Summary: &quot; expr=&quot;summary&quot; &#x2F;&gt; &lt;&#x2F;onentry&gt; &lt;transition target=&quot;done&quot; &#x2F;&gt; </code></pre> &lt;&#x2F;state&gt;<p>&lt;final id=&quot;done&quot; &#x2F;&gt;<p>&lt;&#x2F;agentml&gt;<p>```<p>We’re using this in Agentflare to add observability, cost tracking, and compliance tracing for multi-agent systems — but AgentML itself is fully open-source (MIT licensed).<p>Repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;agentflare-ai&#x2F;agentml" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;agentflare-ai&#x2F;agentml</a> Docs: <a href="https:&#x2F;&#x2F;docs.agentml.dev" rel="nofollow">https:&#x2F;&#x2F;docs.agentml.dev</a><p>We also launched SQLite-Graph, a Cypher-compatible graph extension for SQLite, which will serve as the base for AgentML’s native memory layer. It’s also MIT licensed: <a href="https:&#x2F;&#x2F;github.com&#x2F;agentflare-ai&#x2F;sqlite-graph" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;agentflare-ai&#x2F;sqlite-graph</a><p>Would love feedback from anyone building with LLM orchestration frameworks, rule-based systems, or embedded MCP tool servers… especially around how to extend deterministic patterns to multi-agent coordination.<p>— Jeff @ Agentflare

[Other] Show HN: React-like Declarative DSL for building synthetic LLM datasets

[Monitoring/Observability] Agent-o-rama: build, trace, evaluate, and monitor LLM agents in Java or Clojure

[Other] Mergiraf: Syntax-Aware Merging for Git

[Other] Show HN: a Rust ray tracer that runs on any GPU – even in the browser I’ve been experimenting with Rust lately and wanted a project that would help me explore some of its lower-level and performance-oriented features. Inspired by Sebastian Lague’s videos, I decided to implement my own ray tracer from scratch.<p>The initial goal was just to render a simple 3D scene in the browser at a reasonable frame rate. It evolved into a small renderer that can: • Run locally or on the web using wgpu and WebAssembly • Perform mesh rendering with a Bounding Volume Hierarchy (BVH) for acceleration • Simulate both direct and indirect illumination for photorealistic results • Be deployed easily as a free web demo using GitHub Pages<p>The project is far from perfect, but it’s been a fun way to dig deeper into graphics programming and learn more about Rust’s ecosystem. I’m also planning to experiment with Rust for some ML projects next.<p>GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;tchauffi&#x2F;rust-rasterizer" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;tchauffi&#x2F;rust-rasterizer</a> Web demo (desktop browsers): <a href="https:&#x2F;&#x2F;tchauffi.github.io&#x2F;rust-rasterizer&#x2F;" rel="nofollow">https:&#x2F;&#x2F;tchauffi.github.io&#x2F;rust-rasterizer&#x2F;</a><p>Would love feedback from anyone who’s built similar projects or has experience with wgpu or ray tracing in Rust.

[CLI Tool] Show HN: Serie – A rich Git commit graph in your terminal Serie is a TUI application that uses the terminal emulators&#x27; image display protocol to render commit graphs like git log --graph --all.<p>This is not a full-featured git client, nor are there any plans to add such functionality, so it is not a replacement for tig, lazygit, gitui, etc.<p>The only purpose of this tool is to provide a pretty git log --graph and make commit information easily accessible.<p>While some users prefer to use Git via CLI, they often rely on a GUI or feature-rich TUI to view commit logs. Others may find git log --graph sufficient.<p>Personally, I found the output from git log --graph difficult to read, even with additional options. Learning complex tools just to view logs seemed cumbersome.<p>Limitations:<p>- Sixel is not supported. Only terminals that support the iTerm and kitty image protocols are supported.<p>- Terminal multiplexers are not supported.<p>- Windows is not supported.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;lusingander&#x2F;serie" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;lusingander&#x2F;serie</a>

[Other] Celtic Code: Drawing Knots with Python

[DevOps] Oxy is Cloudflare's Rust-based next generation proxy framework (2023)