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

XY is an extremely fast, interactive, customizable Python charting library for the web, notebooks, and static exports. Charts are composed declaratively or through matplotlib conventions. You can fully customize them with Python, CSS, or Tailwind. With small charts, every point is sent to the browser. For large charts, the Rust core computes only what the screen needs to display, based on its resolution. Pan, zoom, hover, and selection can show full details by running the same process for the new range, and a selection returns the original rows. Important XY is in alpha and is receiving frequent enhancements. ⭐️ Star the repo to follow the progress. XY is for Python users who want one flexible charting library for everything from everyday plots to custom application visuals and large datasets. Build a chart once, then use it in notebooks and web apps or export it as HTML, PNG, SVG, or PDF. pip install xy # or, with uv uv add xy A chart is a container plus the marks inside it. Any sequence works; NumPy is optional. import xy chart = xy.line_chart(xy.line([1, 2, 3, 4, 5], [120, 180, 165, 240, 310])) # chart.to_html("chart.html") # chart.to_png("chart.png") # chart.to_svg("chart.svg") chart # notebooks render it The same API scales to a hundred million points as a density surface: import numpy as np import xy rng = np.random.default_rng(7) n = 100_000_000 r = 6.0 * rng.beta(1.2, 3.0, n) theta = 2.9 * np.log1p(r) + rng.integers(0, 4, n) * (np.pi / 2) + rng.normal(0, 0.045 + 0.016 * r, n) chart = xy.scatter_chart( xy.scatter( r * np.cos(theta), r * np.sin(theta), color=np.exp(-r / 2.2), colormap="magma_r", density=True, opacity=0.85, # Grow and solidify markers once a view drills through to real rows. size=2.5, zoom_size_factor=2.6, zoom_opacity=0.95, ), xy.theme( background="#ffffff", plot_background="#ffffff", grid_color="#e6e6e1", axis_color="#c3c2b7", text_color="#0b0b0b", ), title="100 million points", ) chart For common pyplot workflows, change the import and keep the plotting code: import numpy as np import xy.pyplot as plt x = np.linspace(0, 10, 200) fig, ax = plt.subplots() ax.plot(x, np.sin(x), "r--", label="signal") ax.legend() plt.show() See the compatibility guide; not all charts and functionality are supported yet. Use Python to control the chart, from marks and axes to interactions and layout. - Marks: Control color, size, opacity, symbols, gradients, strokes, curves, and colormaps. - Guides: Customize axes, ticks, grids, annotations, legends, colorbars, and tooltips. - Interaction: Add pan, zoom, hover, selections, crosshairs, callbacks, and linked charts. - Layout: Create layers and facets, set responsive dimensions, and apply themes. chart = xy.line_chart( xy.line(x, y, color="#7c3aed", width=3), class_name="rounded-xl bg-white", class_names={"tooltip": "rounded-lg bg-zinc-900 text-white"}, ) See the styling guide for examples. For a detailed breakdown of what can be customized, see the capability matrix. Live interactive charts, 10k to 100M points. Every library gets every row and is driven through its own input path in a real browser. The clock stops only when the canvas is both correct (planted sentinel points verified lit) and stable (10 byte-identical frames), so progressive renderers are charged until their last chunk lands. XY holds 0.071 s at 10k and 0.081 s at 100M, flat across four orders of magnitude, because above 200k rows it draws a screen-bounded density surface instead of one marker per row, and zoom drills back to exact rows. Every exact-marker path scales with N instead: Matplotlib crosses a second at ~3M and reaches 13.4 s at 50M; Plotly crosses at ~2.5M and reaches 9.8 s at 25M. The pale line is XY with density=False : the same engine drawing one marker per row, no aggregation credit. It renders 100M exact markers in 1.34 s on 5.26 GiB. Time until every point is on screen, in seconds. ✕ is a size the library did not render: Plotly never finishes constructing the figure at 50M, and Matplotlib draws at 100M but never resolves the zoom that follows. | Points | 10k | 100k | 500k | 1M | 2.5M | 5M | 10M | 25M | 50M | 100M | |---|---|---|---|---|---|---|---|---|---|---| | XY speedup | 1× | 2× | 3× | 4× | 9× | 16× | 34× | 89× | 177× | — | | XY | 0.071 | 0.072 | 0.075 | 0.084 | 0.083 | 0.089 | 0.083 | 0.077 | 0.076 | 0.081 | XY (density=False ) | 0.085 | 0.074 | 0.087 | 0.098 | 0.111 | 0.144 | 0.206 | 0.424 | 0.645 | 1.343 | | Matplotlib (WebAgg) | 0.086 | 0.115 | 0.224 | 0.357 | 0.758 | 1.424 | 2.804 | 6.838 | 13.385 | ✕ | | Plotly (scattergl) | 0.341 | 0.373 | 0.477 | 0.614 | 1.033 | 1.785 | 3.367 | 9.794 | ✕ | ✕ | Peak Python-side resident memory, in GiB. Browser memory is tracked separately and excluded here, since a headless Chrome resides ~1 GiB before drawing anything. | Points | 10k | 100k | 500k | 1M | 2.5M | 5M | 10M | 25M | 50M | 100M | |---|---|---|---|---|---|---|---|---|---|---| | XY advantage | 1.8× | 1.7× | 1.9× | 2.1× | 2.1× | 2.4× | 2.6× | 2.9× | 2.8× | — | | XY | 0.05 | 0.05 | 0.06 | 0.07 | 0.13 | 0.19 | 0.32 | 0.70 | 1.36 | 2.58 | XY (density=False ) | 0.05 | 0.05 | 0.07 | 0.10 | 0.18 | 0.31 | 0.57 | 1.35 | 2.66 | 5.26 | | Matplotlib (WebAgg) | 0.09 | 0.09 | 0.12 | 0.15 | 0.28 | 0.46 | 0.84 | 2.06 | 3.85 | ✕ | | Plotly (scattergl) | 0.21 | 0.18 | 0.28 | 0.36 | 0.60 | 1.05 | 1.86 | 4.70 | ✕ | ✕ | One machine (Apple M5 Pro), one run per cell; at the small end the timings carry roughly ±10 ms of run-to-run spread. For the environment, methodology, per-size videos, and raw results, see the benchmark runbook and competitive benchmark specification. The reflex-xy adapter turns any XY chart into a regular Reflex component, with no JavaScript, iframe, or separate chart service. It ships as its own package and pulls in xy and reflex : pip install reflex-xy # or, with uv uv add reflex-xy Register the adapter once: # rxconfig.py import reflex as rx import reflex_xy config = rx.Config( app_name="dashboard", plugins=[reflex_xy.XYPlugin()], ) Then add a chart anywhere in the component tree: import reflex as rx import reflex_xy import xy signups = xy.line_chart( xy.line([1, 2, 3, 4, 5], [120, 180, 165, 240, 310]), title="Weekly signups", ) def index() -> rx.Component: return rx.card( rx.heading("Growth"), reflex_xy.chart(signups, height="320px"), width="100%", ) app = rx.App() app.add_page(index) Hover, pan, and zoom keep working. For charts driven by Reflex state, events, or live streams, see the Reflex integration guide and the runnable example app. Each notebook fetches its rows from the linked public source; no raw datasets are stored in this repository. Counts describe the featured chart, and the notebooks scale further. See the example guide for sources, workload controls, and setup. | Gaia DR3 · HR diagram 250,000 plotted stars Open notebook | gnomAD v4.1 · allele frequency 164,000 plotted variants Open notebook | Pan-UKBB · Manhattan plot 814,294 plotted variants Open notebook | | Dukascopy · EUR/USD ticks 101,427 plotted ticks Open notebook | LIGO · GW150914 strain 16,777,216 raw · 3,441 shown Open notebook | NYC TLC · taxi pickup density 300,000 pickup records Open notebook | Most chart stacks serialize every value as JSON and ask the browser to draw every mark. XY keeps exact values in a ColumnStore , computes a level of detail in Rust, and transfers typed binary buffers. Decimated and density views are bounded by the visible result. flowchart TB API["Python API
Build the chart"] STORE["ColumnStore
Keep canonical f64 columns"] CORE["Native Rust compute
Direct · decimated · density"] PAYLOAD["Compact payload
Data-less JSON spec + typed binary buffers"] RENDER["Browser or notebook
WebGL2 marks · Canvas axes · DOM interface"] API --> STORE --> CORE --> PAYLOAD --> RENDER So a dense overview can aggregate while a narrow view returns exact points. With a live host, pan and zoom request a refined payload. Canonical f64 data stays in Python, so hover and selection still return original rows. For the full design, see the design dossier. Broad 2D coverage first, then geographic, 3D, and volume visualization. Queued next, no dates implied: - Categorical distributions: strip, swarm, beeswarm, boxen, rug - Regression diagnostics: trendline, residual, QQ, PP - Scatter matrix and joint plots: SPLOM, pair grid, marginal histograms - Pie / donut: in xy.pyplot today, promoting toxy.pie_chart(xy.pie(...)) - Candlestick / OHLC and finance overlays: SMA, VWAP, Bollinger, RSI, MACD; prototyped, awaiting a fresh landing - Waterfall and funnel - Treemap, sunburst, and icicle - Radar / polar and gauge: needs polar axes first - Slope, bump, and dumbbell - 3D and volume: scatter, surfaces, meshes, isosurfaces, and volumetric views The full ranked backlog is in the chart roadmap. Want a chart or feature that isn't listed? Open an issue.