Show HN for August 15, 2025

Hi HN,

I’ve been working with Kubernetes for over a decade, since the alpha days, and was involved in kube-aws project before AWS launched EKS. For the past four years, I’ve been helping friends and small businesses cut costs by running Kubernetes on Hetzner Cloud, which I’ve found to be rock solid and by far the best priced provider.

Provisioning a cluster on Hetzner is now straightforward, thanks to tools like k3s and hetzner-k3s, but configuring it for your specific needs still takes time and expertise. I built Edka to make that part easy: spin up a production ready cluster in ~2 minutes, then choose how low level or automated you want to go.

How it works:

Layer 1 – Cluster provisioning - Creates a k3s-based Kubernetes cluster on Hetzner (lightweight, easy to manage, scales well).

Layer 2 – Add-ons - One-click deploy for metrics-server, cert-manager, and various operators; preconfigured for Hetzner, no extra setup needed.

Layer 3 – Applications - Minimal config UIs for apps built on top of add-ons. - Example: Need PostgreSQL? Fill a few fields → platform installs CloudNativePG → provisions HA PostgreSQL with PITR → gives ready to use endpoints. Backups can be restored to any point in time with a click. Quick demo: https://edka.io/apps/

Layer 4 – Deployments - Connect your CI to push container images to a public/private registry. - Edka updates deployments automatically (with semantic versioning rules), supports instant rollbacks, autoscaling, persistent volumes, secrets/env imports, and quick public exposure. Quick demo: https://edka.io/deployments/

Tech stack: TypeScript, React + Tailwind CSS, PostgreSQL, Redis, BullMQ, Vault + AWS KMS to encrypted sensitive data.

The platform is still in beta and I’m building it in my spare time, so there are some rough edges, but I’d love feedback from anyone running Kubernetes on Hetzner, exploring alternatives to EKS/GKE/AKS or looking to automate their infrastructure with Kubernetes.

More details: https://edka.io/

Thank you!

I wrote this JMAP MCP server that adds email management tools to Claude for searching, reading, and sending emails through FastMail and other JMAP providers in Deno!

I needed real-time updates in a web UI whenever PostgreSQL table rows change, so I built PgHook. It's a 23 MB Docker image (10.1 MB compressed), .NET9 AOT-compiled, that streams logical replication events and sends them to a configurable webhook.

In my setup, the webhook converts events to SignalR messages that push updates to the UI.

I know about Debezium but I needed something minimal, and I like playing with csharp.

Clojure Land is a simple website intended to help with discovering open-source Clojure projects.

This is the repository for the Zero Trust VPN code run by VP.NET that allows you to verify that the servers are running the same code as the repository using Intel SGX for hardware guarantees.

I built LDNS because I'm constantly curious about domain names - who owns them, which nameservers they use, how they're configured, and what their DNS records reveal. I wanted a fast, easy way to investigate domains right from my browser without juggling multiple tools.

LDNS runs entirely client-side using Cloudflare's DNS over HTTPS. Just type a domain and instantly see:

All DNS records with clickable filtering RDAP/WHOIS data with registrar info and expiration dates Email security configuration (SPF, DMARC, MTA-STS, BIMI) Export options (JSON, CSV, BIND zones, PDF reports)

Try it at ldns.com - just append any domain like ldns.com/example.com to start investigating.

Built with SvelteKit and deployed on Cloudflare Pages for that instant-load experience we all love.

Comments, feedback, and feature requests welcome!

You can follow along with the progress at the github, but I've added support for Gemma 3 120M, the speed for local LLM to TTS token time is incredible!

The most feature-rich version of Kuvasz since the 2.0.0 release has arrived. Custom status code and keyword matching, slow response checks, new translations, and a lot of smaller improvements and fixes are included in version 2.4.0!

Hello HN!

Between academics and everything else on my plate, I still find myself watching way too many YouTube videos. So I built `youtubegist` - just add `gist` after `youtube` in any video URL to get an instant summary.

Before: https://youtube.com/watch?v=<...> After: https://youtubegist.com/watch?v=<...>

I know there are other YouTube summarization tools, but they're either cluttered, paywalled, or don't format summaries the way I need them. So I made my own that's free, open source, and dead simple.

One cool thing, if you install it as a PWA (on Android using Google Chrome), you can share YouTube URLs into it from the YouTube app, and it should summarize the video for you!

Please leave your feedback if you tried it out! Thank you!

I'm the maintainer of OpenAVMKit. It's a free and open source toolkit for real estate mass appraisal. I want to make it easier for analysts, researchers, and assessors to build automated valuation models (AVMs) from public data sources. It's in early development, but has reached a stage where it's useable and I'm gathering public feedback.

Stuff it can do:

- Data enrichment: add additional data from public sources like OpenStreetMap, Overtrue, US Census, USGS, or your own shapefiles

- Modeling: run many different algorithms, such as MRA, GWR, LightGBM, XGBoost, CatBoost, as well as ensembles, all through one interface

- Reporting: generate IAAO (International Association of Assessing Officers) compatible ratio studies and other statistics and graphs

- Reproducible workflows: build data pipelines out of input data and settings file, so the same person can reproduce your work on another computer starting with the same ingredients

There's a "getting started" section in the docs with a minimal example that should let you download the test jurisdiction and run through the basic example notebooks.

I'm looking for feedback on what features people are the most interested in and what kinds of tutorials/documentation would be the most helpful for me to focus on.

The official site is here: https://www.openavmkit.com

The github is here: https://www.github.com/larsiusprime/openavmkit

We also have a bit of a writeup on it here: https://progressandpoverty.substack.com/p/openavmkit-a-free-...

Happy to answer any questions in the thread.

Hi HN!

I was disappointed when the ChatGPT Agent announcement came with the note that there'd be limited usages available for something that's architecturally simple:

> Pro users have 400 messages per month, while other paid users get 40 messages monthly, with additional usage available via flexible credit-based options.

So assembled this with Cloudflare's recent Containers API. Here's a link to the tweet we posted launching it:

https://x.com/getlsd/status/1956441045119672710

Feel free to fork or star and make funny things happen :)

Hey HN, I am Zuhaib from Magicnode (https://magicnode.ai), no code AI micro apps builder platform visually with workflows. Here is the demo video: https://www.youtube.com/watch?v=pIecK6uKZjU&t=4s&ab_channel=...

With Magicnode, you can:

Build AI apps visually — drag, drop, and connect “nodes” for prompts, logic, APIs, and UI elements.

Publish instantly — your app becomes a shareable web link with a clean interface, no deployment headaches.

Monetize & share — there’s an “App Store” where you can publish your AI tools for others to use, free or paid.

Go beyond chatbots — workflows can handle structured inputs, multi-step logic, file processing, and integrations.

Example: In under 10 minutes, a non-technical marketer made a “viral TikTok caption generator” that uses OpenAI + a trending-hashtag API, and shared it with their whole team.

We’re hoping Magicnode lowers the barrier for creators, small businesses, and indie hackers who want to build AI products fast — without hiring developers or wrangling cloud deployments.

Would love your feedback on:

What use cases you’d want to see

Gaps we should close to make it useful for power users

Features that would make you actually switch from code to visual building

Note: We applied to YC F25 but didn’t get in as of now — still excited to share what we’ve built and get feedback.

We’re live now — try it here: magicnode.ai

Hey HN! I've always been frustrated by having to follow people across 10 different platforms just to catch what they actually say. Lately I've quit following other media entirely and am just hanging out on twitter, but the amount of things I'm missing is vast, I'm sure.

Recently I started playing with this new API from Parallel AI that let's you do super deep web research and collect tons of datapoints at once. Today I tried to solve my own problem and built Based People using the Task API.

*Features:*

- Finds and tracks public appearances – podcasts, conferences, congressional hearings, interviews, fireside chats, and blogs for 160+ influcential tech people - allows you to follow any of them and build a public appearances feed for yourself

It took me a little trial and error and I spent over $200 (should've tested with small test-set longer), but overall I am really impressed how far webscraping has come (they beat GPT5 in their own benchmark). Repo is linked from the footer for more details.

*Current limitations:*

- Only covers 2024-present (historical data was too much for the AI) - Does not refresh automatically yet, will need a daily or weekly update with a cronjob

Show HN: I built a directory of public appearances of tech people across the web

Right now it's free to use, no signup required to browse. It's also open source if you're curious. Next steps are verifying/improving data quality and deepening the dataset.

Would love your feedback on the concept and execution!

Most memory solutions just do RAG search on chat logs. That's not real memory. We want a real memory that could understand a user and grow with them, not just search their past chats.

Memobase works like ChatGPT memory, specially designed for chatbot agents. It's personalized and scales to millions. Retrieval latency is under 80ms. ( 85% score on LOCOMO Temporal Benchmark https://github.com/memodb-io/memobase/tree/main/docs/experim...

- Profile + Timeline schema — not complex embedding. - Context API — full system prompts for evolving & personalized LLM responses. - Built with FastAPI, Postgres, Redis — ready for production.

Check it out on GitHub: https://github.com/memodb-io/memobase

I'm tired of using the keyboard that is there on the Android.

The voice part is not very good, so I am thinking from scratch what should be the voice native keyboard looks like.

I just launched the app. It is in beta. Anyone interested to try it out? Let me know or reach out to me on rahul[@]dictationdaddy.com

Free. Self-Hosted. Open Source Collectors. Bring Your Own LLM.

I built TimeBrain since I wasn't satisfied with existing activity trackers. I wanted to understand what I spent my day working on, while remaining in control of my data, and without spending hours trying to decipher crowded graphs.

You can try the demo with test data: https://demo.timebrain.io If you are interested in participating in the beta I will be releasing invite codes to download the full version on the Discord, which you can find on the site.

Please let me know any feedback or your opinions on my product :)

I’ve shipped a big performance upgrade to Sarpro, an open‑source Rust tool for converting Sentinel‑1 SAR GRD products into GeoTIFF or JPEG. Since my initial post [here](https://news.ycombinator.com/item?id=44872768), Sarpro now avoids full‑resolution I/O when you only need a smaller output and performs reprojection in‑process without writing giant temporary files.

Highlights: - Target‑size reads and single‑step warp: read/warp directly to the final output size instead of loading full‑res first. This cuts I/O and memory by 5–20× for small outputs. - Reprojection to any CRS: `--target-crs EPSG:xxxx` with in‑process gdalwarp via VRT (no temp GeoTIFF). Resampling: nearest/bilinear/cubic/lanczos. - Faster autoscaling: O(N) histogram‑based percentiles replace O(N log N) sorting. - Batch + GUI: both now honor reprojection/resampling in batch mode. - Performance: on a modern laptop (M4Pro12), scaling a dual‑band ~400–500MP GRD to 2048 px typically completes in ~1–2 s; no‑warp downsamples can be sub‑second. Full native warps remain tens of seconds as expected.

Features: - CLI, GUI, and Rust API - Synthetic RGB from polarization pairs, robust autoscaling, optional padding - TIFF and JPEG outputs with georeferencing/sidecars, metadata emission

I’d love feedback from the RS/EO community: more RGB presets, additional processing modes, tiling for ML, and cloud pipeline integrations.

Links: - GitHub: `https://github.com/bogwi/sarpro` - Previous HN thread: [Show HN post](https://news.ycombinator.com/item?id=44872768)

Hi HN, I'm Neha and I built April to manage my email & calendar while driving.

April is a voice assistant that helps manage my emails & calendar schedule while I drive or walk. I can just say: - “Reply to Sasha” and dictate the message - April sends the email right away.

- “Archive these emails” “file this to the Clients folder,” or “delete all promos” - all hands-free.

- Schedule or reschedule meetings without touching my screen.

Now, in 20 minutes of commute, I’m at Inbox Zero and my calendar is set for the day. I walk into work already prepared for the day.

This one change freed up a full hour every morning. It buys you back time if you're buried under email and have a long commute to work.

I'd like your feedback on - your most painful email & calendar workflows

- what other integrations you'd like to have a voice interface for

- what would make April useful for you