2026年4月17日 の Show HN

I built MCP servers for my oscilloscope and SPICE simulator so Claude Code can close the loop between simulation and real hardware.

We built AI Subroutines in rtrvr.ai. Record a browser task once, save it as a callable tool, replay it at: zero token cost, zero LLM inference delay, and zero mistakes.

The subroutine itself is a deterministic script composed of discovered network calls hitting the site's backend as well as page interactions like click/type/find.

The key architectural decision: the script executes inside the webpage itself, not through a proxy, not in a headless worker, not out of process. The script dispatches requests from the tab's execution context, so auth, CSRF, TLS session, and signed headers get added to all requests and propagate for free. No certificate installation, no TLS fingerprint modification, no separate auth stack to maintain.

During recording, the extension intercepts network requests (MAIN-world fetch/XHR patch + webRequest fallback). We score and trim ~300 requests down to ~5 based on method, timing relative to DOM events, and origin. Volatile GraphQL operation IDs are detected and force a DOM-only fallback before they break silently on the next run.

The generated code combines network calls with DOM actions (click, type, find) in the same function via an rtrvr.* helper namespace. Point the agent at a spreadsheet of 500 rows and with just one LLM call parameters are assigned and 500 Subroutines kicked off.

Key use cases:

- record sending IG DM, then have reusable and callable routine to send DMs at zero token cost

- create routine getting latest products in site catalog, call it to get thousands of products via direct graphql queries

- setup routine to file EHR form based on parameters to the tool, AI infers parameters from current page context and calls tool

- reuse routine daily to sync outbound messages on LinkedIn/Slack/Gmail to a CRM using a MCP server

We see the fundamental reason that browser agents haven't taken off is that for repetitive tasks going through the inference loop is unnecessary. Better to just record once, and get the LLM to generate a script leveraging all the possible ways to interact with a site and the wider web like directly calling backed API's, interacting with the DOM, and calling 3P tools/APIs/MCP servers.

I've been presenting at local meetups about Context Engineering, RAG, Skills, etc.. I even have a vbrownbag coming up on LinkedIn about this topic so I figured I would make a basic example that uses bedrock so I can use it in my talks or vbrownbags. Hopefully it's useful.

I’ve been working on Pyra for the past few months and wanted to start sharing it in public.

Right now it’s focused on the core package/project management workflow: Python installs, init, add/remove, lockfiles, env sync, and running commands in the managed env.

The bigger thing I’m exploring is whether Python could eventually support a more cohesive toolchain story overall, more in the direction of Bun: not just packaging, but maybe over time testing, tasks, notebooks, and other common workflow tools feeling like one system instead of a bunch of separate pieces.

It’s still early, and I’m definitely not claiming it’s as mature as uv. I’m mostly sharing it now because I want honest feedback on whether the direction feels interesting or misguided.

I couldn't find a database that was immutable (like Datomic) and embedded (like SQLite), so I made one. The reference impl is in Zig and there are now ports for Java, Clojure, TypeScript, and Go.

Waputer is an operating system that runs entirely in the browser. When you visit the website at https://waputer.app, a kernel written in JavaScript sets up a filesystem and launches a WebAssembly program, which in turn talks to the kernel to handle the display and input. A purely terminal-based version is at https://waputer.dev.

My original intention was to create programs that run in the browser that have a lot more in common with the desktop. The traditional "hello world" program is not really suited for the web. Waputer changes that. The GitHub repo at https://github.com/waputer/docs gives a very brief overview of compiling a C program and running it on Waputer. There is a blog available from the main site that has a long-form explanation of Waputer and my motivations if you want some additional reading.

I made this tool that allows server admins to decrypt passwords during bootup with the help of Matrix and HTTP.

Built this around one simple idea:

the workflow that wants to execute should not be the same place that decides whether execution may continue.

This project puts an external allow/deny boundary before action.

Public entry points:

* live pilot * commercial request * private deployment

There is also a GitHub Marketplace action install surface, but the main point is the boundary model itself: decision stays outside the workflow that is asking to proceed.

Looking for feedback from people working on CI/CD, security controls, approval boundaries, and automated execution.

Lately I've been working on a small tool to generate realistic handwriting animation from any text. Any feedback is welcome, hope you enjoy :)

Start with a FREE instant website for your AI on the open internet, then work with it to build a business that sells specialized datasets, files, premium reports, blogs, courses and more.

Hi HN,

I built a small macOS utility to solve something that kept bothering me: external drives not being safely ejected when a Mac goes to sleep, leading to “Disk Not Ejected Properly” warnings and potential data issues.

macOS doesn’t always give volumes enough time to unmount, especially with SSDs, SD cards, or disk images. Manually ejecting works, but it’s easy to forget.

Ejectify automates this by unmounting selected volumes right before sleep (or display off) and mounting them again on wake.

In Ejectify 2, I focused on making (un)mounting more reliable, with an optional helper for better consistency, improved sleep handling to give operations more time to complete, broader volume support, and the ability to suppress system warnings if they still occur.

To celebrate the launch, it’s available for €4.99 (instead of €6.99) with the code EJECTIFY2.

Best, Niels.

Small cli program I made to convert and modify bookmark files. Supports converting between json and netscape bookmark file format (default formatted exported by chrome/firefox).

I created this because I have a lot of bookmarks across devices that I want to batch edit/delete and I can't always just directly modify the local browser db.

Not many filters implemented so far, but I made it easy to add filters see: https://github.com/ediw8311xht/cl-bookmark-tool/blob/main/sr...

A living collection of real prompts shared by creators on Twitter, covering data viz, portraits, product design, and more.Links back to original tweets.