Show HN for May 5, 2026

I built PaletteInspiration.com, a browsable archive of color palettes pulled from artworks by 3,000+ master painters (Monet, Vermeer, Raphael, Van Gogh). Why I built it: every color palette generator I tried converged on the same five muted pastels. Painters spent centuries figuring out color and we mostly ignore that body of work when picking colors for digital design. Please share your feedback on the Color Harmony Explorer - drag the wheel to any color and it shows which hues master painters historically paired with it (not only standard complementary, analogous, triadic, etc.) It is solely based on co-occurrence across thousands of real paintings. Not algorithmic color theory rules - actual empirical pairings.

No signup, no paywall, no email capture. Just curious what people think.

I’m Michel, co-founder and CEO of Airbyte (https://airbyte.com/). We’ve spent the last six years building data connectors. Today we're launching Airbyte Agents (https://docs.airbyte.com/ai-agents/), a unified data layer for agents to discover information and take action across operational systems.

Here’s a quick walkthrough: https://www.youtube.com/watch?v=ZosDytyf1fg

As agents move into real workflows, they need access to more tools (e.g. Slack, Salesforce, Linear). That means a ton of API plumbing: authentication, pagination, filters, handling schema, and matching entities across systems.

Most MCPs don’t fix this. They’re thin wrappers over APIs, so agents inherit their weak primitives and still get it wrong most of the time, especially when working across tools.

An even deeper issue is that APIs assume you already know what to query (think endpoints, Object IDs, fields), whereas agents usually start one step earlier: they need first to discover what matters before they can even start reasoning.

So we built Airbyte Agents to be a context layer between your Agents and all of your data. The core of this is something we call Context Store: a data index optimized for agentic search, populated by our replication connectors. All that work on data connectors the last six years comes in handy here!

This gives agents a structured way to discover data, while still allowing them to read and write directly to the upstream system when needed.

What got us working on this was an insane trace from an agent we were migrating to our new SDK. It was supposed to answer "which customers are at risk of leaving this quarter?" The trace had 47 steps. Most were API calls. The agent first had to find a bunch of accounts, then map them to the right customers, then look for tickets, bla bla... and when the Agent finally responded, the answer sounded ok, but was wrong. Not only that, it was excruciatingly slow. So we had to do something about it.

That 47-step agent is one example of a question where Airbyte Agents does particularly well. Other examples: - “Show me all enterprise deals closing this month with open support tickets." - “Find every support ticket that doesn’t have a Github issue opened”

Some of these might sound simple, but the quality of the answer changes dramatically when the agent doesn’t have to assemble all that context at runtime.

Once we had an early version of the product, I spent a weekend building a benchmark harness to see if it worked. Also for fun, I like writing benchmarks :). I compared calling the Airbyte Agent MCP vs calling a bunch of vendor MCPs directly. I tested retrieval, and search.

For the sake of simplicity, I used token consumption as a unit of measure. I think that’s a good proxy for how well agents are working. A failing agent (like the one that took 47 steps), will churn through lots of tokens while getting nowhere, while a successful one will get straight to the point.

Here's what I found when measuring: for Gong, it used up to 80% fewer tokens than their own MCP, for Zendesk up to 90% fewer, for Linear up to 75%, and for Salesforce up to 16% (Salesforce’s own SOQL does a good job here).

Of course there is the usual obvious bias: we are the builders of what we are benchmarking. So we made the test harness public: https://github.com/airbytehq/airbyte-agents-benchmarks. Feel free to poke at it, and please tell us what you find if you do!

It's still early and some parts are rough, but we wanted to share this with the community asap. We'd love to hear from people building agents: - Are you indexing data ahead of time, or letting the agent call APIs live? - How are you matching entities across systems?

Would also love to hear any thoughts, comments, or ideas of how we could make this better, and if there are obvious things we’re missing. For now, we’re excited to keep building!

Hi HN! Looking for feedback on this 1v1 and daily word dueling game I've built over the last few months.

Play here: https://wordtrak.com/

Or on iOS here: https://apps.apple.com/us/app/wordtrak/id6760442363 (Android version soon!)

I've going through a backlog of projects and checking them off with Claude Code. What started out as a quick afternoon of coding with claude has relight my love of making.

As we all are, I've been experimenting with ways to reduce external saas spend, and continually bring traditionally external pieces of context (prs, docs, trello boards) into the one mono repo.

I have toyed with a markdown todo list and setting claude/codex into a "ralph" loop previously, but always missed the web ui for organising, planning, and tracking things in progress.

So codex and I build kanban-cli, an installable or npx-runnable cli util that takes one or more markdown files, and uses markdown headers and todo syntax to track and manage agentic tasks.

Let me know your thoughts, or if you've got a more convenient way of managing tasks!

I was tempted to link to the project landing page https://kanban-cli.vochsel.com as I made a cute example demo, but linking to the repo felt less self-promo-y.

Cheers!

Furwall is a tiny macOS menu bar app. While you're at the keyboard or mouse, the FaceTime camera looks for a human face or upper body. When it doesn't find one, the keyboard stops accepting input. Cat walks across your laptop, nothing happens to your code.

Some notes:

Apple's Vision framework runs locally. Video is processed in memory and never uploaded. On a block, Furwall saves one local JPEG to ~/.furwall/catpures/. A second Vision pass throws out anything that isn't a cat, so the daily count in the menu only reflects confirmed cats. There is now a folder on my disk that is slowly filling up with photos of Pepper and Beets walking across my keyboard.

The camera turns on only while you're at the computer (typing, mouse motion, app switch, screen wake) and powers down 30 seconds after the last activity. The green camera dot tracks that.

The keystroke drop uses a CGEventTap at .defaultTap. Furwall ships unsandboxed because of this. A .listenOnly tap with Input Monitoring is enough to see keys, but dropping them needs .defaultTap, which needs Accessibility, which the App Sandbox blocks. Watching keystrokes is sandbox-compatible; stopping them is not.

Mouse events are observed (to wake the camera) but never intercepted or dropped, so the menu bar always works. Three escape hatches: click the icon and quit, mash Escape five times in 1.5 seconds for a 5-minute pause, or revoke Accessibility in System Settings (macOS invalidates the tap). If Vision stalls for any reason the keyboard fails open after 10 seconds, which is better than soft-bricking the machine.

Furwall never uploads camera frames or keystrokes. Its own network traffic is Sparkle update checks plus the donate sheet's anonymous totals/click counter. One short charity slug per click, no user identifier.

The donate item in the menu opens the donate page of a vetted animal-welfare charity for your system Region. Ten orgs across nine regions: Alley Cat Allies and PetSmart Charities in the US, Cats Protection in the UK, Cat Protection Society NSW in Australia, Toronto Cat Rescue in Canada, NSPCA in Ireland, SPCA in New Zealand, Deutscher Tierschutzbund in Germany, La SPA and Fondation 30 Millions d'Amis in France, Japan SPCA in Japan. Each org is registered or recognized under its local charity or nonprofit regime, and the list gets re-vetted every release. No money flows through the app.

macOS 15+, signed and notarized, MIT.

https://olliewagner.com/furwall

It started with drones. My co-founder and I were building a Minecraft plugin where drones had to navigate through unloaded chunks - terrain that doesn't even exist in memory yet. No existing library had any concept of it, which led to drones being stuck in non-existence.

Sloppy at first, tightly coupled to Minecraft's API, we built our own. But somewhere along the way we realized: the problem wasn't Minecraft. It was how Java pathfinding is built in general - object-heavy, GC-hostile, and single-threaded by assumption.

So I split the project. The core became its own thing: pure Java 8+ for compatibility, featuring a zero-allocation primitive heap instead of a theoretically optimal FibonacciHeap, because pointer chasing and cache misses were unacceptable. 100% async, even in a single-threaded, async-hating Minecraft hell, and a processor pipeline providing full extensibility. No game engine, yet the possibility to adapt to your environment.

Sometimes it isn't the algorithm that needs optimization, but the environment it lives in. Pathetic is full of these micro-optimizations: a BloomFilter lookup before the expensive closed set check, a multi-metric squared heuristic to kill Math.sqrt() overhead, and bit-packing 3D coordinates into a single primitive long. In order to get Java fast, you need to treat it like C++.

The result: 10k concurrent paths in ~7ms, where most libraries quietly die. A 20k distance path (equivalent to 100k+ nodes) in ~60ms without bloating the heap. The README is... a choice. Judge it, because it does the same to you.

Cryptographically secure and verifiably robust protection with drop-in integration and outcome-based pricing

AI-DLC-UML modifies AI-DLC to enable AI agents to drive the software development workflow with UML modeling. It is intended for those who want to use UML modeling collaboratively in their design practices, even in AI-driven software development.

Hi everyone- 've spent a lot of time recently working on this AI-powered song conversion web app. I think it's ready for testing. You can search for songs, link to a song on YouTube, Spotify, Tidal, etc, or even upload a file from your own machine, and my service will convert it quickly and in top quality.

Have a look, try searching for a song, and let me know if you can suggest any improvements to our service.

Keyterm prompting is a valuable way to help your STT better recognize unique terms like brand names etc, but for non-English languages/non-standard accents, providers like Deepgram tend to hallucinate keyterms in STT transcripts. So the output transcript contains the given keyterms, even when those keyterms are not present in the input audio.

I'm currently collecting feedback to improve this product. Right now it cuts down keyterm hallucinations by about 60% on in-house test data, so I'm curious to see how it performs in public.

The product is free to use while in beta (Hindi and Indian-accented English are supported). Would love to hear how it performs on your data. Feel free to drop a comment if you’re interested in features like additional language support, streaming and self-hosting.

We are all here because we love HN. The site itself, well simple is beautiful. But what if it could feel more alive? What if you no longer had to refresh it several times a day. What if you could just leave the site open and have stories automatically refresh and visually move when rank changes? What if it had a ticker showing new posts alongside the top posts?

I built RNKFlow.com (Rank Flow) to make that happen. Using the official HN API, the site will feel familiar yet different. Leave the site up for a bit and watch the magic happen. Stories move position, new comments are highlighted, top risers stand out, and new stories scroll along the footer.

One of the many cool features... Categories

Only want to see AI or Programming related stories? Select the various filters below the header to see what YOU want to see. The filter you choose also applies to both the top story and the new story ticker feed.

Free tier is just that, free for everyone, no account required. See up to 60 posts at a time, Dark mode available, filter one category at a time. If you choose to sign in, you can save your theme, story view count and row density preferences.

Did I mention the mobile web view looks great?

Pro tier ($5) opens up some cool new features:

Stories - Push beyond the default 60 stories to 500. Never miss a thing. Categories - Select multiple categories at a time. The feed will blend the chosen categories into one feed Alerts - Have a business and want to keep tabs on your domain name? Have a topic that really interests you? All of that and more can be monitored with alerts. Bookmarks - Easily bookmark stories to view later Export - Want to export you bookmarked stories or alerts? CSV and JSON export built in.

Give a try. I really hope you enjoy the experience and would love to hear your feedback.

It's a clone of 7 Little Words, but with topics from computer science and programming. No sign-up, no app install, no tracking. It's a PWA and works offline, so you can add it to your home screen if you'd like.

Tech: js, no libs, Canvas API, Web Audio, not vibe coded, but I did use Claude for graphics and tests, puzzles curated by hand.

I made an interactive visualizer for Hugging Face models!

Paste any Hugging Face url to view the model in multiple granularities.

Would love to hear feedback and feature requests!

I’ve spent a few months building this. It’s basically a visual env for the Intel 4004 because I wanted to see what's actually happening in the buses and registers without 'black boxes'.

It has a transpiler called QuadBasic to go from high-level syntaxs to native MCS-4 machine code. Just a personal project from a senior dev to tribute the 1971 pioneers. Happy to answer any tech questions if you find it interesting.

We've just (re-)launched Arta Labs! Join our subreddit for a give-to-get feedback system. Before you can post your idea, you must review 4 other ones first. This provides a better incentive structure than the take-take environments many other forums tend to be.

Additionally, we have also set up a private community for aspiring or existing entrepreneurs working on ideas, big or small, to build alongside each other. We organise co-working days, private group chat, a knowledge lab, are also building a mentor network where members can book 1-1 calls with experts.

Find out more at https://artalabs.com or watch this:

https://youtube.com/watch?v=uAbey_7HUQM&t=1s

I just built this with a basic set of guardrails and a limited skillset, and it’s still pretty rough around the edges.

I’d really appreciate any feedback.

It’s free for the first few users — you can use the invite code DEMO34.

Note: This is experimental and not fully tested. It can and will make mistakes, so use it carefully, especially in important or sensitive email threads.