MCP server MatPlotLibNet

MatPlotLibNet — matplotlib for .NET

MatPlotLibNet is a charting library for C#, .NET 10 and .NET 8, inspired by matplotlib. It draws 91 chart types and saves them as SVG, PNG, PDF or an animated GIF, straight from your server code. It does not need a JavaScript framework, a WebView or a hosted service.

It has a fluent API and it works with dependency injection. It ships with 148 colormaps, 30 themes, 13 map projections with embedded Natural Earth data, LaTeX-style MathText, text in any script, O(1) streaming (57 of its 59 technical indicators have a streaming form), a control room (Plt.OpsDashboard() — KPI tiles, state timelines, one shared trend window), and an MCP server so an AI agent can draw with it. Charts render natively in Blazor, WPF, MAUI, Avalonia, Uno and ASP.NET Core, inside a Verso notebook cell, and TypeScript clients are there for Angular, React and Vue.

Where this is going

1.29.0 puts shapes on an axes and lets an axis show another unit. A rectangle, a polygon, an arrow, a circle, an ellipse, a wedge and an arc (AddPatch) are placed in data coordinates and bend with the axes they stand on, as matplotlib's patches do: on a log axis a circle bends and a straight edge stays straight. They count in the automatic range, a labelled one has a row in the legend, and they travel through JSON and the MCP server. A top or side axis can show the same values in another unit through a function and its inverse (Link with an AxisTransform: degrees Celsius against Fahrenheit, a frequency against its period), as matplotlib's secondary_xaxis and secondary_yaxis do. The top X axis now follows its own tick locator and scale, a series on it follows a log primary Y axis, a translucent line or text keeps its opacity in the SVG, a figure built with AddSubPlot(rows, cols, index) keeps its cell through JSON, and the MCP server draws a spec that holds shapes, spans or annotations and no series. The three rows that were still to come on the roadmap are done.

1.28.0 draws more of what matplotlib draws. Stairs, VLines, HLines and Triplot join the chart verbs, and a sixth family of verbs (Psd, Csd, Cohere, MagnitudeSpectrum, PhaseSpectrum, AngleSpectrum, Acorr, Xcorr) draws a sampled signal in the frequency domain, and its correlation, with matplotlib's defaults; the twelve bring the chart verbs to thirty-eight. A contour label follows its line, upright and in the line's colour; a figure carries text, a legend and an image of its own (FigText, FigLegend, FigImage), drawn on a figure with no subplot too; and an inset marks the region it zooms into (IndicateZoom). The spectrogram windows each frame once and draws its rows at the frequencies they are, a line that is downsampled keeps its gaps, and a line set to LineStyle.None draws no line. A figure's texts, legend and images and an inset's zoom travel through JSON and the MCP server.

1.27.0 finishes the roadmap. A word cloud, a chord diagram, a ternary plot and a Piper diagram are new charts, a selection in one chart can light the same rows in another, and an animation can be saved as one SVG that steps through its frames. A live chart sends its own gestures, one handler interface receives them, and several viewers of one chart share it by design. The native controls return Home to the first view, the MAUI view reaches the same functions, and the Windows heads are tested by CI. 57 of the 59 indicators have a streaming form that follows each new bar, and transfer entropy follows a window.

1.26.0 gives marks that are a size in data units a home. A waffle chart, a calendar of events, Point and Figure and Renko charts draw squares, stripes and boxes that tile the plot area, and Kagi draws its thick and thin lines. A fill can run from one colour to another, an arrow can follow a connected scatter, and the frame of a plot can be trimmed: a spine spans only the data's range, a plot keeps an equal scale on request, and parallel coordinates label every column in its own units. An icicle chart, an arc diagram and a stem-and-leaf plot join the layouts, a waterfall chart can mark a target on each bar, and the comparison with matplotlib now runs in CI. Three new series types (the waffle chart, Point and Figure and the icicle chart) join the chart types, and six new verbs (CalendarEvents, Renko, Kagi, ConnectedScatter, StemAndLeaf, MatShow) bring the chart verbs to 26.

1.25.0 gives each value a chart resolves one owner that every reader asks. The colour bar puts its ticks where the chart's normaliser puts its colours, a theme owns the neutral grey that reference lines and outlines wear, a grid the author changed stays a change over whichever theme the figure is drawn under (a polar chart's rings and a 3-D chart's walls included), and a light's strength reaches the surfaces and bars it lights. Saving to .png, .pdf or .gif without the Skia package is refused rather than written as SVG, and axes that share a scale share its margins, so a table stands directly under the curve it counts. The statistics grow: the violin reads the same Gaussian estimator as the density plot, ProbabilityPlot draws a Q-Q plot against an exponential, uniform, log-normal or Student t distribution (or one of your own), RiskTable puts the numbers at risk and the log-rank test under a survival plot, and ShewhartChart draws the individuals, moving range, X-bar, range, p and c charts and marks points by Nelson's eight tests. The volume tools gain an anchored VWAP (Vwap(bars, anchor)) and a volume profile (VolumeProfile). The library now has 59 indicators.

1.24.0 gathers every Y axis of a chart into one container. The primary axis is the first axis of the left side, and each further axis stacks outward past the one before it, on either side and as many times as the chart needs one: WithAddRightYAxis and WithAddLeftYAxis add an axis with its own range, ticks, colour, series and legend entry, there is no "second" or "third" in the name to count, and the four scales 1.23.0 stopped at are no limit any more. The older verbs for a secondary and for an outward-stacked Y axis keep working and are marked obsolete in favour of them; the changelog lists each with what replaces it. A Y axis title now stands the same distance from its tick labels on the left and on the right, 4 px between the title and the widest label by default, and WithYAxisTitleGap sets another distance for the whole chart; the left title no longer runs off the edge of the figure. Every axis draws the way the primary one does, with its own log or symlog scale, tick locator, minor ticks and a title in math markup, and a line on a left-hand axis can carry its name at its end. A figure written to JSON keeps every axis, whatever their number, and a figure that stays within the old four writes the file 1.23.0 wrote. A chart server drops a zoom or pan that names a subplot it does not have, and one event that fails no longer ends the chart's session for good, as it used to; the MCP server checks every axis of a chart spec instead of the first two. The texts of a chart are placed with more care: a legend left at Best stands where it covers the fewest marks, as matplotlib's loc='best' does, an annotation takes the theme's text size and can move by a number of pixels or out of the way of the data and the legend, a bar's value label stands beyond the bar's end whichever way the bar points, and a turned tick label has room beside the plot.

The same release picks up thirteen more rows of the roadmap, each one rule where there had been two. A colour bar puts its ticks at round numbers, as many as fit along its length. A polar chart has one radial scale that its rings, lines, bars and heatmap all read, and polar bars stack, so a wind rose is one PolarBar call for each speed class. Bars stack on both sides of zero, so a Likert scale is a stacked bar chart. A filled contour fills the bands between its levels over the grid and follows the limits you set. A series that cannot be drawn from the data it was given is refused when it is built, and not when the chart is drawn, and the MCP server names the place in the spec where it stands. A stacked area names its layers in the legend, a state timeline's legend entry shows its states, and a series says for itself whether it needs axes around it. A map from MatPlotLibNet.Geo keeps its holes, comes back from JSON, colours its regions by a value (Choropleth) and takes points, bubbles, hexagon bins, flows along great circles, a cartogram and a Dorling map.

1.23.0 moves the technical indicators that read a candle's open, high, low, close and volume from separate array parameters to one OhlcvBars value, built with the factory that names the columns the indicator actually uses (OhlcvBars.Hlc(high, low, close), OhlcvBars.Cv(close, volume), and so on) instead of repeating the same arrays across every indicator on a chart. A chart with more than two scales now has somewhere to put the extra ones: a verb that stacks a second axis past the first on either side — four scales total, two left and two right, each with its own range, ticks, series and legend entry, the way matplotlib moves a twin's spine outward. A secondary axis restores through the same registry the primary one does, so a bar or an area survives there and on the top axis too, not only a line or a point; and a wheel zoom, a drag pan, a rectangle zoom and Home now move every one of those axes, each by its own amount, whether the server answers the gesture or a native control does. Underneath, the close to a hundred chart verbs the fluent API builds from now live one-per-file as extension members on Axes instead of all sharing one class — a file-layout change only, no call a reader writes is any different because of it.

1.22.0 lets several series share one colour instead of each taking the next one from the theme's cycle, and spends it on three things: a Kaplan-Meier curve now draws its confidence band and its censor marks in the curve's own colour, and two new verbs, Raincloud (a half violin, a narrow box and a group's raw points in one colour) and Horizon (a signed series folded into bands of two hues), bring the axes builder to seventeen verbs.

1.21.0 makes every chart type and every axes setting round-trip through JSON whole. A Sankey's nodes and links, a treemap's, sunburst's, dendrogram's and clustermap's trees, a tree grid's rows, a polar chart's data, a streaming chart's current points and a planar 3-D bar's data all come back from JSON as themselves now, not a placeholder — the MCP server lists and accepts all eight again, where it used to refuse them by name. An axes' Z axis, its 3-D pane colours and its tooltip switch come back too. The one series this release did not reach is MatPlotLibNet.Geo's GeoPolygonSeries, named on the roadmap; 1.24.0 brings it back from JSON too.

1.20.0 turns charts that took a page of code into one call. Fifteen verbs on the axes builder, among them Pareto, Ridgeline, CalendarHeatmap, QQPlot, ControlChart, KaplanMeier, StreamGraph and Marimekko, take the data as you hold it (names, groups, samples, dates) and build the chart from the series the library already has. Underneath them, a figure written to JSON now reads back as the picture it was, with every setting it draws, so the MCP server and every transport carry it whole. NaN is the one missing value: a line breaks at a gap, as in matplotlib, and nothing is drawn at NaN. Bars on a log axis reach the bottom edge, horizontal bars work, category names are tick labels with room for the axis title, and WithDpi and SaveOptions now reach the file. Every new call is shown in code on the wiki's examples page.

1.19.0 is for charts with more than one line and more than one axis. WithDirectLabel() writes a line's name at its end, in the line's own colour, instead of in the legend, and SetYAxisColor / SetXAxisColor colour an axis to match its line in one call, on the left, bottom, right and top axis alike. The right-hand and top axes now follow their own tick settings, get room of their own in the layout and keep their colours through JSON, and a line on them hides when its legend entry is clicked. On a server-interactive chart and in the WPF, Avalonia and Uno controls, hover, selection, zoom, pan and the data cursor report the values the reader sees, on log axes too, and an axis limit you set, reversed or not, is now exactly the range drawn. Every new call is shown in code on the wiki's examples page, and every C# example in the wiki now compiles.

After that, releases follow what the community needs: bug fixes, documentation, and whatever comes up in real use. Open a Discussion or an Issue; that is what decides the next release. Every release, with its migration notes, is listed in the CHANGELOG, and the roadmap says what is done, what is still to come and what is left out on purpose, with a section for each row on the wiki's page.

Quality bar: every class must pass a strict coverage gate of ≥90 % line and branch coverage (1,161 classes, at 99.8 % line / 98.9 % branch) across 28,839 tests. A separate suite (Tst/MatPlotLibNet.Fidelity) compares the rendering with reference images drawn by matplotlib, and CI runs it.

---

Documentation

• Wiki — getting started, cheatsheet, all 91 chart types, the control room, styling, streaming, SignalR, packages
• From matplotlib to C# — the pyplot call you know, and the line that draws it here
• Cookbook — copy-paste examples with rendered output
• Playground — try charts live in the browser
• API Reference — generated from the XML docs
• Roadmap — what is done and what is still to come, in the order each row needs the one above it; the wiki's page says what each row takes and what is left out on purpose
• Benchmarks · Coverage policy · CHANGELOG

---

Packages

Package	Install	What it does
MatPlotLibNet	dotnet add package MatPlotLibNet	Core: models, fluent API, SVG rendering, JSON, transforms
MatPlotLibNet.DataFrame	dotnet add package MatPlotLibNet.DataFrame	Microsoft.Data.Analysis.DataFrame extension methods — plot, indicators (SMA/EMA/RSI/MACD/…), and polynomial regression from named columns
MatPlotLibNet.Skia	dotnet add package MatPlotLibNet.Skia	PNG, PDF, and animated GIF export via SkiaSharp
MatPlotLibNet.Blazor	dotnet add package MatPlotLibNet.Blazor	MplChart + MplLiveChart Razor components with SignalR
MatPlotLibNet.AspNetCore	dotnet add package MatPlotLibNet.AspNetCore	REST endpoints, SignalR hub, IChartPublisher
MatPlotLibNet.Interactive	dotnet add package MatPlotLibNet.Interactive	figure.ShowAsync() — browser popup, no server needed
MatPlotLibNet.GraphQL	dotnet add package MatPlotLibNet.GraphQL	GraphQL queries + subscriptions via HotChocolate
MatPlotLibNet.Maui	dotnet add package MatPlotLibNet.Maui	Native MplChartView via Microsoft.Maui.Graphics
MatPlotLibNet.Avalonia	dotnet add package MatPlotLibNet.Avalonia	Native MplChartControl for Avalonia 12 — Skia backend, optional local interaction
MatPlotLibNet.Uno	dotnet add package MatPlotLibNet.Uno	Native MplChartElement for Uno Platform (WinUI 3 / Android / iOS / macCatalyst)
MatPlotLibNet.Wpf	dotnet add package MatPlotLibNet.Wpf	Native WPF MplChartControl via SkiaSharp — all 15 interaction modifiers — and MplStreamingChartControl for live data; your project targets netX.0-windows10.0.19041.0
MatPlotLibNet.Geo	dotnet add package MatPlotLibNet.Geo	13 map projections, GeoJSON parser, Natural Earth 110m data, geographic polygons
MatPlotLibNet.Verso	Verso → Extensions panel	A chart inside a Verso notebook cell, drawn as SVG (cookbook)
MatPlotLibNet.Notebooks	#r "nuget: MatPlotLibNet.Notebooks"	Inline SVG in Polyglot / Jupyter notebooks. Microsoft has ended Polyglot Notebooks and .NET Interactive, which this package builds on, so existing notebooks keep working but nothing new is coming — for new work use MatPlotLibNet.Verso
MatPlotLibNet.Mcp	dnx MatPlotLibNet.Mcp	MCP server — an AI agent renders charts over stdio (cookbook)
@matplotlibnet/angular	npm install @matplotlibnet/angular	Angular components + TypeScript SignalR client
@matplotlibnet/react	npm install @matplotlibnet/react	React hooks + components + TypeScript SignalR client
@matplotlibnet/vue	npm install @matplotlibnet/vue	Vue 3 composables + TypeScript SignalR client

---

Quick start

using MatPlotLibNet;
using MatPlotLibNet.Styling;

double[] x = [1, 2, 3, 4, 5];
double[] y = [2, 4, 3, 5, 1];

Plt.Create()
    .WithTitle("My First Chart")
    .WithTheme(Theme.Dark)
    .AddSubPlot(1, 1, 1, ax => ax
        .Plot(x, y, s => { s.Color = Colors.Blue; s.Label = "Data"; })
        .WithLegend())
    .Save("chart.svg");

---