Show HN for September 23, 2025

I built a tool called *Kekkai* for file integrity monitoring in production environments. It records file hashes during deployment and later verifies them to detect unauthorized modifications (e.g. from OS command injection or tampering).

Why it matters:

* Many web apps (PHP, Ruby, Python, etc.) on AWS EC2 need a lightweight way to confirm their code hasn’t been changed. * Traditional approaches that rely on metadata often create false positives. * Kekkai checks only file content, so it reliably detects real changes. * I’ve deployed it to an EC2 PHP application in production, and it’s working smoothly so far.

Key points:

* *Content-only hashing* (ignores timestamps/metadata) * *Symlink protection* (detects swaps/changes) * *Secure S3 storage* (deploy servers write-only, app servers read-only) * *Single Go binary* with minimal dependencies

Would love feedback from others running apps on EC2 or managing file integrity in production.

Hey HN! A few months ago we shared our AI dataset generator as an open source repo, and the response was incredible (https://news.ycombinator.com/item?id=44388093). We got requests from folks who wanted to use it without the hosting overhead, so we created both options: a hosted version (https://www.metabase.com/ai-data-generator for instant use and the source code fully open (https://github.com/metabase/dataset-generator) for anyone who wants to self-host or contribute.

Looking forward to seeing how you use it and what you build on top of it!

Bonus: The repo now supports multi-provider LLM integration with LiteLLM, thanks to a great contribution from their team.

This is a project I made over the weekend. What if when you SSH'd into a machine, instead of giving you a user login, it would provision a new microVM just for you?

It's like a little SimCity for VMs.

This turned out to be a lot of fun, and I've wanted to make my own custom SSH server for a while now, with its fancy progress bars and color animations. The hardest part was definitely debugging VM setup, boot, and networking though! Compiling the Linux kernel with all the right configs was difficult. For a while I had issues with silent hangs on boot due to lack of available system entropy, on Linux <4.19. Anyway, I wrote a bit about my struggles. :D

I managed to bundle everything (vmlinux 6.1, firecracker, SSH server, iptables/bridge/masquerade) into a single, statically-linked binary for x86_64/aarch64. It takes the kernel rootfs as a command-line argument.

We built FlyCode after seeing subscription businesses lose ~35% of recurring revenue each year to failed payments — even when customers had other valid cards on file.

*The problem:* When a customer's primary card fails, Stripe retries a few times then cancels the subscription. If that customer has a backup card, it isn’t tried. At least 20% of active customers have more than one card on file, which means a lot of preventable churn.

*Our solution:* FlyCode automatically identifies if a customer has other valid cards on file and retries them when a subscription payment fails. You can configure when these retries happen during the dunning period (beginning, middle, end) and define validity rules (e.g. “card was used in last 180 days”). It’s a Stripe app — no code changes needed.

We've seen 18%-20% higher recovery rates from our core retry engine, plus another 5–10% from using backup cards. Importantly, there was no increase in refunds or chargebacks — in fact, rates were lower than merchant averages. Big companies like Microsoft and Amazon already do this internally; we wanted to make the same capability accessible to smaller SaaS teams.

*Under the hood:* FlyCode monitors for failed invoices, checks for available backup methods via Stripe’s PaymentMethod API, and systematically retries in a way that avoids service disruption or manual workflows.

We’re Jake, Etai, and Tzachi — we previously built payment recovery systems at startups and enterprises, which is how we discovered this gap.

You can try it here: [https://www.flycode.com/stripe]

We’d love feedback from anyone dealing with subscription payment failures. What’s been your experience with involuntary churn? Have you considered leveraging backup payment methods?

Hi guys, I'm sharing something I've been tinkering on for a while: https://signagesync.app/

Like Google Chromecast, but "cast" to multiple screens at once. Create a playlist of auto-refreshing web pages, videos, and even live streams (but not supported on Windows). Also works with local network, e.g. http://192.168.../sales-dashboard

It's still early (MVP), but already usable—and I'd love to hear your feedback.

Tech stack: SvelteKit, WebSocket, Flutter (desktop)

I’m a student founder, and Snapdeck has been a 6-month sprint from idea to product. In March I hacked together a minimal version and launched it on Product Hunt just to see if anyone else had the same frustration with slides — and people did. A couple of months later I tried a Figma plugin version, which brought in more feedback but still felt limited.

Yesterday we launched the first full standalone version on Product Hunt and reached #2 of the day. About 1,000 people have signed up so far, with ~30% retention among the main target users.

Under the hood, Snapdeck runs on an orchestration layer we built that routes tasks across multiple open-source language models and commercial APIs, acting as an “agent” to generate structured slides and editable charts. Unlike most AI slide tools, everything stays fully editable — layouts, visuals, and content can be dragged, modified, or updated via natural-language commands. Current output is PDF, with PPTX coming soon. Demo screenshots: https://youtu.be/fmlQ6cccj1w .

We’re considering open-sourcing parts of the orchestration layer, and I’d love feedback from HN: is this genuinely useful for real workflows, and what’s still missing?

https://snapdeck.site

Hey everyone! I'm excited to unveil WindowSill, a universal command bar for Windows 10 and 11 that brings AI-powered text assistance and a suite of productivity tools directly to wherever you're working.

Why WindowSill? Inspired by the MacBook Touch Bar, StreamDeck, and Apple Intelligence, I wanted a tool that gives context-aware actions without interrupting my workflow. WindowSill is my take on that idea for Windows.

What it can do

AI text assistant Select any text in any app to summarize, rewrite, translate, or fix grammar instantly. No copy/paste needed. No switching apps needed.

Short-Term Reminders Set reminders that can pop up as full-screen notifications, so you can't miss them. Perfect for ADHD brains, multitaskers, or anyone who needs a firm nudge to stay on track of things.

Clipboard history Access your recent copies without switching apps.

URL & text utilities Select any URL in any app to shorten it or generate a QR Code.

Media & Meetings controls Manage playback, mute/unmute from Microsoft Teams, even when the app is in the background or minimized.

Personalization Save custom prompts, dock the "sill" to the top, bottom, left, right, or change its size to reflect your style and needs.

All from a single, universal command bar that stays out of your way — no friction, no app switching.

Bonus: the app is extensible! It comes with an SDK allowing any app to integrate with WindowSill.

Who is it for? Mainly Windows power users, but also anyone looking to boost their productivity with AI-powered text assistance and quick access to useful tools.

Try it today for free! Visit https://getwindowsill.app Product Hunt launch: https://www.producthunt.com/products/windowsill-2

I'd love your feedback: what features would make WindowSill most useful for you? Or what would you like to see next?

I've been frustrated with dependency hell and clutter on my VPS from dev, so I built Devbox: a lightweight, open-source CLI tool that spins up isolated development environments using Docker. Each project runs in its own container, but your code stays in simple flat folders on the host machine—no messing with volumes or sync issues. Environments are disposable, so you can nuke and recreate them without losing your work. Key features:

- Instant setup: `devbox init my-project` and you're in a fresh env with `devbox shell`.

- Configurable via JSON: Define packages, services, and more in a `devbox.json` file. Share it in your repo for reproducible setups—teammates just run `devbox up`.

- Docker-in-Docker by default: Build and run containers inside your env without extra config.

- Host-friendly: Edit code directly on your machine; the container handles the runtime.

- Templates for quick starts: Built-ins for Python, Node.js, Go, web dev, etc.

- Advanced options: Port mapping, env vars, resource limits, and even mounting your dotfiles.

It's FOSS (MIT license), Linux-focused (Debian/Ubuntu, or WSL2 on Windows), and super easy to install: `curl -fsSL https://devbox.ar0.eu/install.sh | bash`.

Check out the launch page and docs at https://devbox.ar0.eu, or the repo at https://github.com/itzCozi/devbox. I'd love some feedback, stars, or contributions to help grow this into a solid community tool!

Upload photos from multiple angles for fast, accurate mushroom identification with critical safety information, toxicity warnings, and similar species alerts.

Personal learning project. Just looking for feedback on getting better at C and CUDA in general.

I had this idea: what if you removed customer choice entirely?

So I'm selling rocks for $49.99. You can't pick which one. You just get rock #000001, then the next person gets #000002, etc. No returns, no exchanges.

Currently sourcing rocks and taking pre-orders for November. Each one gets weighed, photographed, numbered, and comes with a certificate. Could be a 1-gram pebble or a 10kg boulder - same price.

It's the opposite of how everything online works. Amazon shows you a million options, I'm giving you zero. Wondering if that constraint makes it more interesting or if I'm just making it harder for no reason.

No social media, no marketing, barely any explanation on the site. Just: here's a rock, here's a number, here's the price.

Honestly not sure if anyone will actually buy these, but the pre-orders will tell me if removing choice kills demand or creates it.

https://weight.rocks

Hey HN,

I just want to share Forklet, a tool we built out of frustration with GitHub's all-or-nothing approach to repository access.

The Problem: Our security team needed to scan specific files in pull requests (using our scanner Valkyrie), but we kept hitting two walls:

1. Cloning entire multi-GB repos just to check a few config files

2. GitHub API's rate limits making targeted downloads painful

The Solution: Forklet gives you surgical precision for GitHub downloads:

# Download only Python files, excluding tests forklet download owner/repo ./target --include "*/.py" --exclude "test_"

# Or just grab a specific folder forklet download owner/repo ./configs --target-paths "src/config"

Looking forward to your feedback and ideas!

Mirrow is a small TypeScript DSL that compiles to SVG. It aims to make SVG and animation simpler without JSX or hand rolled CSS.

Highlights:

- Compile time checks and attribute validation - Inline events like on:click and @hover - CLI to turn .mirrow files into static SVGs or use as components

Quick try: npx mirrow -i example.mirrow -o example.svg

Playground: [https://mirrow.app/playground](https://mirrow.app/playground) GitHub: [https://github.com/MirrowApp/mirrow](https://github.com/MirrowApp/mirrow)

Feedback on DX, docs, and rough edges is very welcome. Thank you

RaceThere.com was designed to be simple. You set the location and filters of race events you're looking for, and it generates a full list of upcoming events in that area. It's as easy as that to find your next race.

Sure there are other websites out there like RaceThere.com, but who wants to scroll through a website that hasn't been touched since 2010.

Do you value websites that look nice more?

Hello HN, a quick share from my weekend project. TBX Live Server is an extension that bakes a browser-like webview right into VS Code, keeps multiple servers in sync, and reloads on the fly. Developers can run parallel environments, toggle ports per workspace, and stay in the editor while testing. In the last 24 hours it picked up 19 new downloads and 2 stars, so early adopters seem to be finding it useful.

Repo: http://github.com/TheMailmans/vscode-inline-live-server Marketplace install: https://marketplace.visualstudio.com/items?itemName=Thinkbac... Happy to answer questions or hear what workflows you’d like to see supported next.

Hey HN, I built this simple tool for fun over the weekend after getting tired of breaking my flow to copy and paste what I was reading in a Claude tab.

My goal was to make the process as frictionless as possible so you don't expend cognitive load thinking about the tool. To that end, there are no hotkeys or buttons to initiate the chat widget, the extension just detects natural language as you type and populates the widget after a threshold.

The LLM gets the text content in your current viewport as context.

This is a minimal distraction free journalling app. Features include

- only one entry per day - sync across cloud - pwa - no likes, no comments, no validations and hence no distraction

Got this idea to develop a habit of writing everyday. Welcome to all feedback and suggestions.

P.S: This is a very basic mvp built using cursor, next.js and supabase

This started from a chat with my brother-in-law. He was starting a new business and planning to spend a lot of time on SEO. As i’d just started using ChatGPT more often myself, I asked him whether he was optimising for LLM answers, not just Google? Whilst Semrush and Ahrefs were leaning into this space, there was no dedicated tool looking at this.

Anyway, after a few months of building I have now created Surchee:

- Crawl your site and see how AI engines (ChatGPT, Perplexity, Claude, Bing Copilot, etc.) summarise it - Check if your brand, value proposition, or trust signals are coming through clearly - Track whether AI bots are even visiting your site (via a small pixel) - Use dashboards that look a bit like SEO tools, but aimed at LLM/AI visibility instead

It’s quite fresh but the idea is simple: type in a URL, get a score and some insights.

If anyone has any general feedback or more specific feedback per the items below, it would be appreciated:

- What metrics would make this more useful - Whether “AI visibility” feels like a problem worth solving - Any rough edges or confusing parts in the current version

Link: https://www.surchee.com

Happy to answer any questions too.

Domain Watch emails you when a domain name is listed for sale on an aftermarket or its price drops.

It's like camelcamelcamel.com for domain names.

Currently supports the following "Buy It Now" aftermarkets:

- Afternic

- Sedo

- Spaceship SellerHub

- Atom

- Namecheap Market

- Porkbun Marketplace

- Gname

You can watch up to 3 domains for 60 days with the free plan.

Support for auctions is coming soon.

Why I made this:

1. To know where and when I could buy the domains that I wanted to buy, ASAP. 2. Empower other domain name buyers with marketplace awareness, data, and an improved shopping experience. 3. To make money one day.

Feedback welcome and appreciated!

Hi HN,

A few weeks ago I shared Kvatch.com a federated SQL engine I’ve been hacking on that lets you query across multiple data sources with plain SQL.

Since then, I’ve added support for querying live APIs alongside files and databases. That means you can now join data from:

REST APIs

CSV files

Google Sheets

Postgres / SQLite

Git repositories

All as if they were one datasource.

Example use cases:

Join your GitHub issues with commits from a repo

Combine a Google Sheet of leads with enrichment data from a REST API

Build quick dashboards that mix API + database + CSV data

Kvatch is open-source, written in Go, and still early — but I’d love feedback on what data sources / features would make it most useful.

Docs & example -> https://github.com/kvatch-hub/kvatch-cli/tree/main/examples

GitHub -> https://github.com/kvatch-hub/kvatch-cli

Website -> http://kvatch.com

Blog -> http://kvatch.com/blog

Thanks for checking it out!