Show HN: XY – Fast, composable, GPU-accelerated charts, written in Rust

This new chart app has coders cheering, side-eyeing, and begging to escape old tools

TLDR: XY is a new Python chart tool promising blazing-fast visuals, even with huge piles of data, while still working in notebooks, websites, and exports. Commenters were split between hype over escaping older tools and skepticism that most people need this much speed at all.

A new project called XY just rolled into Show HN promising very fast charts, smooth zooming, and the ability to handle everything from tiny notebook graphs to absolutely gigantic datasets. The big flex? It’s built so regular Python users can make one chart and reuse it on the web, in notebooks, or export it as images and documents. It’s also proudly in alpha, which in normal-person terms means: exciting, ambitious, and very much still being assembled in public.

But let’s be honest: the real fireworks were in the comments. One camp was instantly smitten, with people practically swooning over the combo of Rust and Tailwind, a sentence that somehow reads like catnip for modern developer culture. Another commenter summed up years of pain in academia with a deadpan cry for freedom from matplotlib, the old plotting tool many researchers use because, well, that’s what everyone already uses. That gave the thread a faint "new king arrives to challenge the dusty empire" energy.

Then came the skeptics. The hottest pushback was basically: do we really need GPU speed for charts? One commenter argued that most charts aren’t big enough for this to matter, and if they are, they may already be too cluttered to understand. Ouch. Another user delivered a classic practical gut-punch: "no support for native GUI?" In other words, cool demo, but can it fit into real workflows? So the vibe was equal parts hype, hope, and hard side-eye — exactly the kind of messy launch-day drama the internet lives for.

Key Points

  • XY is an alpha-stage Python charting library with a Rust core for interactive charts on the web, in notebooks, and in static exports.
  • For small charts, XY sends all points to the browser; for large charts, it computes only screen-needed output based on display resolution.
  • The library supports both declarative chart composition and matplotlib-style workflows through an `xy.pyplot` compatibility layer.
  • The same API is presented as scaling from simple line charts to density-based rendering of datasets with 100 million points.
  • The article claims benchmark performance of about 0.071 seconds at 10k points and 0.081 seconds at 100M points by using screen-bounded density rendering above 200k rows.

Hottest takes

"Rust and Tailwind CSS two of my favorite things in one" — farhan99
"been stuck on matplotlib for centuries" — NickyHeC
"I’m not convinced GPU acceleration is a meaningful advantage" — kasts
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.