The best Hacker News stories from Show from the past day

from

All Show

from the past*

day week month

with these search terms:

Create follow Go back

Latest posts:

Show HN: BunkerWeb – the open-source and cloud-native WAF

Show HN: I AI-coded a tower defense game and documented the whole process

I'm a software developer with 20+ years of experience but during this time I never programmed any games, but I really wanted to for the longest time. With the advent of AI coding agents I thought that this is the best time to try and so I've learned a bit of Phaser.js (a Javascript based game engine) and entered Beginner's Jam Summer 2025 - a game jam for beginners in the game dev industry that allows AI coding. After around 25-30 hours (working mainly after my full-time day job) I managed to submit the game I called "Tower of Time" (the theme of the jam was "Time Travel").<p>You can play it in your browser here: <a href="https://m4v3k.itch.io/tower-of-time" rel="nofollow">https://m4v3k.itch.io/tower-of-time</a><p>The goal of this project for me was first and foremost to see if AI coding is good enough to help me with creating something that's actually fun to play and to my delight is turns out the answer is yes! I decided to document the whole process for myself and others to learn from my mistakes, so both the code AND all the prompts I used are published on GitHub (see submission link). The art assets are largely taken from itch.io artists who shared them for free, with some slight touch ups. Sounds came from freesound.org.<p>I've also streamed parts of the process, you can watch me working on the final stretch and submitting the finished game (warning, it's 5+ hours long):<p><a href="https://www.twitch.tv/videos/2503428478" rel="nofollow">https://www.twitch.tv/videos/2503428478</a><p>During this process I've learned a lot and I want to use this knowledge in my next project that will hopefully be more ambitious. If you have any comments or questions I'm here to answer!

Show HN: I AI-coded a tower defense game and documented the whole process

I'm a software developer with 20+ years of experience but during this time I never programmed any games, but I really wanted to for the longest time. With the advent of AI coding agents I thought that this is the best time to try and so I've learned a bit of Phaser.js (a Javascript based game engine) and entered Beginner's Jam Summer 2025 - a game jam for beginners in the game dev industry that allows AI coding. After around 25-30 hours (working mainly after my full-time day job) I managed to submit the game I called "Tower of Time" (the theme of the jam was "Time Travel").<p>You can play it in your browser here: <a href="https://m4v3k.itch.io/tower-of-time" rel="nofollow">https://m4v3k.itch.io/tower-of-time</a><p>The goal of this project for me was first and foremost to see if AI coding is good enough to help me with creating something that's actually fun to play and to my delight is turns out the answer is yes! I decided to document the whole process for myself and others to learn from my mistakes, so both the code AND all the prompts I used are published on GitHub (see submission link). The art assets are largely taken from itch.io artists who shared them for free, with some slight touch ups. Sounds came from freesound.org.<p>I've also streamed parts of the process, you can watch me working on the final stretch and submitting the finished game (warning, it's 5+ hours long):<p><a href="https://www.twitch.tv/videos/2503428478" rel="nofollow">https://www.twitch.tv/videos/2503428478</a><p>During this process I've learned a lot and I want to use this knowledge in my next project that will hopefully be more ambitious. If you have any comments or questions I'm here to answer!

Show HN: I AI-coded a tower defense game and documented the whole process

I'm a software developer with 20+ years of experience but during this time I never programmed any games, but I really wanted to for the longest time. With the advent of AI coding agents I thought that this is the best time to try and so I've learned a bit of Phaser.js (a Javascript based game engine) and entered Beginner's Jam Summer 2025 - a game jam for beginners in the game dev industry that allows AI coding. After around 25-30 hours (working mainly after my full-time day job) I managed to submit the game I called "Tower of Time" (the theme of the jam was "Time Travel").<p>You can play it in your browser here: <a href="https://m4v3k.itch.io/tower-of-time" rel="nofollow">https://m4v3k.itch.io/tower-of-time</a><p>The goal of this project for me was first and foremost to see if AI coding is good enough to help me with creating something that's actually fun to play and to my delight is turns out the answer is yes! I decided to document the whole process for myself and others to learn from my mistakes, so both the code AND all the prompts I used are published on GitHub (see submission link). The art assets are largely taken from itch.io artists who shared them for free, with some slight touch ups. Sounds came from freesound.org.<p>I've also streamed parts of the process, you can watch me working on the final stretch and submitting the finished game (warning, it's 5+ hours long):<p><a href="https://www.twitch.tv/videos/2503428478" rel="nofollow">https://www.twitch.tv/videos/2503428478</a><p>During this process I've learned a lot and I want to use this knowledge in my next project that will hopefully be more ambitious. If you have any comments or questions I'm here to answer!

Show HN: I rewrote my notepad calculator as a local-first app with CRDT syncing

I launched NumPad v1 on here a few years ago, and back then it wasn't much more than a thin CodeMirror wrapper around the calculator engine I'd written.<p>Now I've rewritten it as a PWA that supports multiple documents, persists them to IndexedDB, and has a syncing service for paying customers. Syncing is handled by Automerge[1] under the hood, which <i>should</i> make it relatively easy to get document sharing working too.<p>[1] <a href="https://automerge.org/" rel="nofollow">https://automerge.org/</a>

Show HN: I rewrote my notepad calculator as a local-first app with CRDT syncing

I launched NumPad v1 on here a few years ago, and back then it wasn't much more than a thin CodeMirror wrapper around the calculator engine I'd written.<p>Now I've rewritten it as a PWA that supports multiple documents, persists them to IndexedDB, and has a syncing service for paying customers. Syncing is handled by Automerge[1] under the hood, which <i>should</i> make it relatively easy to get document sharing working too.<p>[1] <a href="https://automerge.org/" rel="nofollow">https://automerge.org/</a>

Show HN: HomeBrew HN – Generate personal context for content ranking

TLDR: Build a quick HN profile to see how little context LLMs need to personalise your feed. Rate 30 posts once, get a permanent ranked homepage you can return to.<p>Our goal was to build a tool that allowed us to test a range of "personal contexts" on a very focused everyday use case for us, reading HN!<p>We are exploring use of personal context with LLMs, specifically what kind of data, how much, and with how much additional effort on the user’s part was needed to get decent results. The test tool was a bit of fun on its own so we re-skinned it and decided to post it here.<p>First time posting anything on HN but folks at work encouraged me to drop a link. Keen on feedback or other interesting projects thinking about bootstrapping personal context for LLM workflows!

Show HN: HomeBrew HN – Generate personal context for content ranking

TLDR: Build a quick HN profile to see how little context LLMs need to personalise your feed. Rate 30 posts once, get a permanent ranked homepage you can return to.<p>Our goal was to build a tool that allowed us to test a range of "personal contexts" on a very focused everyday use case for us, reading HN!<p>We are exploring use of personal context with LLMs, specifically what kind of data, how much, and with how much additional effort on the user’s part was needed to get decent results. The test tool was a bit of fun on its own so we re-skinned it and decided to post it here.<p>First time posting anything on HN but folks at work encouraged me to drop a link. Keen on feedback or other interesting projects thinking about bootstrapping personal context for LLM workflows!

Show HN: HomeBrew HN – Generate personal context for content ranking

TLDR: Build a quick HN profile to see how little context LLMs need to personalise your feed. Rate 30 posts once, get a permanent ranked homepage you can return to.<p>Our goal was to build a tool that allowed us to test a range of "personal contexts" on a very focused everyday use case for us, reading HN!<p>We are exploring use of personal context with LLMs, specifically what kind of data, how much, and with how much additional effort on the user’s part was needed to get decent results. The test tool was a bit of fun on its own so we re-skinned it and decided to post it here.<p>First time posting anything on HN but folks at work encouraged me to drop a link. Keen on feedback or other interesting projects thinking about bootstrapping personal context for LLM workflows!

Show HN: A modern C++20 AI SDK (GPT‑4o, Claude 3.5, tool‑calling)

Hi all,<p>I’m hacking on new features for the ClickHouse native client and wanted the same “just call the model” ergonomics JavaScript and Python now enjoy. It didn’t exist for modern C++, so I wrote one.<p>ai‑sdk‑cpp (Apache‑2.0) gives you:<p>- Unified calls to OpenAI (GPT‑4o) and Anthropic (Claude 3.5) with a single C++20 API. - Streaming, multi‑turn chat, error handling—all std::optional/std::variant, no macros. - Tool calling (function‑calling) so the model can hit real APIs; sync or async, runs in parallel.<p>The tricky bit: C++ still lacks real reflection, so mapping plain functions → JSON schemas isn’t as automatic as, say, TypeScript decorators. I’d love fresh eyes on that part. Try the examples and tell me where it feels clunky. This is inspired by Vercel's AI SDK [1], and litellm [2].<p>Repo live here: <a href="https://github.com/ClickHouse/ai-sdk-cpp">https://github.com/ClickHouse/ai-sdk-cpp</a>, feedback welcome!<p>[1] <a href="https://github.com/vercel/ai">https://github.com/vercel/ai</a> [2] <a href="https://github.com/BerriAI/litellm">https://github.com/BerriAI/litellm</a>

Show HN: A modern C++20 AI SDK (GPT‑4o, Claude 3.5, tool‑calling)

Hi all,<p>I’m hacking on new features for the ClickHouse native client and wanted the same “just call the model” ergonomics JavaScript and Python now enjoy. It didn’t exist for modern C++, so I wrote one.<p>ai‑sdk‑cpp (Apache‑2.0) gives you:<p>- Unified calls to OpenAI (GPT‑4o) and Anthropic (Claude 3.5) with a single C++20 API. - Streaming, multi‑turn chat, error handling—all std::optional/std::variant, no macros. - Tool calling (function‑calling) so the model can hit real APIs; sync or async, runs in parallel.<p>The tricky bit: C++ still lacks real reflection, so mapping plain functions → JSON schemas isn’t as automatic as, say, TypeScript decorators. I’d love fresh eyes on that part. Try the examples and tell me where it feels clunky. This is inspired by Vercel's AI SDK [1], and litellm [2].<p>Repo live here: <a href="https://github.com/ClickHouse/ai-sdk-cpp">https://github.com/ClickHouse/ai-sdk-cpp</a>, feedback welcome!<p>[1] <a href="https://github.com/vercel/ai">https://github.com/vercel/ai</a> [2] <a href="https://github.com/BerriAI/litellm">https://github.com/BerriAI/litellm</a>

Show HN: Arch-Router – 1.5B model for LLM routing by preferences, not benchmarks

Hi HN — we're the team behind Arch (<a href="https://github.com/katanemo/archgw">https://github.com/katanemo/archgw</a>), an open-source proxy for LLMs written in Rust. Today we're releasing Arch-Router (<a href="https://huggingface.co/katanemo/Arch-Router-1.5B" rel="nofollow">https://huggingface.co/katanemo/Arch-Router-1.5B</a>), a 1.5B router model for preference-based routing, now integrated into the proxy. As teams integrate multiple LLMs - each with different strengths, styles, or cost/latency profiles — routing the right prompt to the right model becomes a critical part of the application design. But it's still an open problem. Most routing systems fall into two camps:<p>- Embedding-based routers use intent classifiers — label a prompt as “support,” “SQL,” or “math,” then route to a matching model. This works for simple tasks but breaks down in real conversations. Users shift topics mid-conversation, task boundaries blur, and product changes require retraining classifiers.<p>- Performance-based routers pick models based on benchmarks like MMLU or MT-Bench, or based on latency or cost curves. But benchmarks often miss what matters in production: domain-specific quality or subjective preferences like “Will legal accept this clause?”<p>Arch-Router takes a different approach: route by preferences written in plain language. You write rules like “contract clauses → GPT-4o” or “quick travel tips → Gemini Flash.” The router maps the prompt (and conversation context) to those rules using a lightweight 1.5B autoregressive model. No retraining, no fragile if/else chains. We built this with input from teams at Twilio and Atlassian. It handles intent drift, supports multi-turn conversations, and lets you swap in or out models with a one-line change to the routing policy. Full details are in our paper (<a href="https://arxiv.org/abs/2506.16655" rel="nofollow">https://arxiv.org/abs/2506.16655</a>), but here's a snapshot:<p>Specs:<p>- 1.5B params — runs on a single GPU (or CPU for testing)<p>- No retraining needed — point it at any mix of LLMs<p>- Cost and latency aware — route heavy tasks to expensive models, light tasks to faster/cheaper ones<p>- Outperforms larger closed models on our conversational routing benchmarks (details in the paper)<p>Links:<p>- Arch Proxy (open source): <a href="https://github.com/katanemo/archgw">https://github.com/katanemo/archgw</a><p>- Model + code: <a href="https://huggingface.co/katanemo/Arch-Router-1.5B" rel="nofollow">https://huggingface.co/katanemo/Arch-Router-1.5B</a><p>- Paper: <a href="https://arxiv.org/abs/2506.16655" rel="nofollow">https://arxiv.org/abs/2506.16655</a>

Show HN: I made a 2D game engine in Dart

Show HN: I made a 2D game engine in Dart

Show HN: CSS generator for a high-def glass effect

There are lots of glassmorphism generators out there, but I wanted to push the effect further! This project is the result of months of experimenting with CSS property layering and battling browser quirks.<p>Cross-browser compatibility is actually the reason I rely on ::before and ::after pseudo-elements to build up the effect. Move the color/opacity to the main element, and you’ll get weird color bleed on the corners in Chrome. Move the texture, and it muddles the bevel’s specular highlight. Move the bevel, and it gets blurred out by the backdrop-filter. And so on!<p>Layers include: * Adjustable blur, brightness, and saturation (backdrop-filter) * Subtle translucent texture * Faux 3D bevel (using box-shadows, not an outline)<p>Glassmorphism is rather heavy on resources, so it’s best used as an accent and avoided on wide desktop elements. Should be compatible with recent versions of Chrome, Safari, and Firefox (desktop and mobile). If you spot bugs or rendering glitches, I’d love to know!<p>Side note: this is an early preview of a framework-agnostic glass SCSS/component library I’m building.

Show HN: CSS generator for a high-def glass effect

There are lots of glassmorphism generators out there, but I wanted to push the effect further! This project is the result of months of experimenting with CSS property layering and battling browser quirks.<p>Cross-browser compatibility is actually the reason I rely on ::before and ::after pseudo-elements to build up the effect. Move the color/opacity to the main element, and you’ll get weird color bleed on the corners in Chrome. Move the texture, and it muddles the bevel’s specular highlight. Move the bevel, and it gets blurred out by the backdrop-filter. And so on!<p>Layers include: * Adjustable blur, brightness, and saturation (backdrop-filter) * Subtle translucent texture * Faux 3D bevel (using box-shadows, not an outline)<p>Glassmorphism is rather heavy on resources, so it’s best used as an accent and avoided on wide desktop elements. Should be compatible with recent versions of Chrome, Safari, and Firefox (desktop and mobile). If you spot bugs or rendering glitches, I’d love to know!<p>Side note: this is an early preview of a framework-agnostic glass SCSS/component library I’m building.

Show HN: CSS generator for a high-def glass effect

There are lots of glassmorphism generators out there, but I wanted to push the effect further! This project is the result of months of experimenting with CSS property layering and battling browser quirks.<p>Cross-browser compatibility is actually the reason I rely on ::before and ::after pseudo-elements to build up the effect. Move the color/opacity to the main element, and you’ll get weird color bleed on the corners in Chrome. Move the texture, and it muddles the bevel’s specular highlight. Move the bevel, and it gets blurred out by the backdrop-filter. And so on!<p>Layers include: * Adjustable blur, brightness, and saturation (backdrop-filter) * Subtle translucent texture * Faux 3D bevel (using box-shadows, not an outline)<p>Glassmorphism is rather heavy on resources, so it’s best used as an accent and avoided on wide desktop elements. Should be compatible with recent versions of Chrome, Safari, and Firefox (desktop and mobile). If you spot bugs or rendering glitches, I’d love to know!<p>Side note: this is an early preview of a framework-agnostic glass SCSS/component library I’m building.

Show HN: HackerNewt – Breadth-first exploring HN client for iOS

I made a HN client that allows you to explore comments in a breadth-first manner. Compared to the classic depth-first approach, it works much better for the larger threads - you don't lose the context and can easily skip discussions you're not interested in.<p>Video preview: <a href="https://imgur.com/a/tzBdpXw" rel="nofollow">https://imgur.com/a/tzBdpXw</a>

Show HN: HackerNewt – Breadth-first exploring HN client for iOS

I made a HN client that allows you to explore comments in a breadth-first manner. Compared to the classic depth-first approach, it works much better for the larger threads - you don't lose the context and can easily skip discussions you're not interested in.<p>Video preview: <a href="https://imgur.com/a/tzBdpXw" rel="nofollow">https://imgur.com/a/tzBdpXw</a>

Show HN: Core – open source memory graph for LLMs – shareable, user owned

I keep running in the same problem of each AI app “remembers” me in its own silo. ChatGPT knows my project details, Cursor forgets them, Claude starts from zero… so I end up re-explaining myself dozens of times a day across these apps.<p>The deeper problem<p>1. Not portable – context is vendor-locked; nothing travels across tools.<p>2. Not relational – most memory systems store only the latest fact (“sticky notes”) with no history or provenance.<p>3. Not yours – your AI memory is sensitive first-party data, yet you have no control over where it lives or how it’s queried.<p>Demo video: <a href="https://youtu.be/iANZ32dnK60" rel="nofollow">https://youtu.be/iANZ32dnK60</a><p>Repo: <a href="https://github.com/RedPlanetHQ/core">https://github.com/RedPlanetHQ/core</a><p>What we built<p>- CORE (Context Oriented Relational Engine): An open source, shareable knowledge graph (your memory vault) that lets any LLM (ChatGPT, Cursor, Claude, SOL, etc.) share and query the same persistent context.<p>- Temporal + relational: Every fact gets a full version history (who, when, why), and nothing is wiped out when you change it—just timestamped and retired.<p>- Local-first or hosted: Run it offline in Docker, or use our hosted instance. You choose which memories sync and which stay private.<p>Try it<p>- Hosted free tier (HN launch): <a href="https://core.heysol.ai">https://core.heysol.ai</a><p>- Docs: <a href="https://docs.heysol.ai/core/overview">https://docs.heysol.ai/core/overview</a>