ChartCS

Chart.js-compatible chart rendering for .NET — straight to PNG, no browser required.

ChartCS ports Chart.js semantics (option names, defaults, scale math, spline math, colors) to C# and rasterizes charts server-side with SkiaSharp. If you know Chart.js, you already know ChartCS — it even accepts your existing Chart.js JSON configs.

Features

• 🎯 8 chart types — bar (grouped & stacked), line, pie, doughnut, radar, polar area, scatter, bubble
• 🧩 Chart.js JSON in, PNG out — feed it real Chart.js configuration JSON (type / data / options)
• 🧮 Faithful to Chart.js — nice-number ticks, beginAtZero/suggestedMin/max semantics, bezier & monotone interpolation, spanGaps, stacking, legends per dataset or per slice
• 🖥️ Server-side & headless — SkiaSharp rendering; no Chromium, no Node, no JS engine
• 🏗️ Three small packages — ChartCS.Core (model + JSON), ChartCS.SkiaSharp (renderer), ChartCS.Fluent (builder API)

Install

dotnet add package ChartCS.SkiaSharp # renderer (pulls in ChartCS.Core)
dotnet add package ChartCS.Fluent # optional fluent builder

Linux note: SkiaSharp ships Windows/macOS natives in-box. On Linux (servers, Docker, CI) also add SkiaSharp.NativeAssets.Linux.

Quick start

usingChartCS.Core;usingChartCS.SkiaSharp;varconfig=newChartConfig{Type=ChartType.Bar,Data=newChartData{Labels=["Jan","Feb","Mar","Apr","May","Jun","Jul"],Datasets=[new(){Label="Sales 2024",Data=[65,59,80,81,56,55,72],BackgroundColors=["#4dc9f6"],BorderWidth=1},new(){Label="Sales 2025",Data=[28,48,40,19,86,27,90],BackgroundColors=["#f67019"],BorderWidth=1}]},Options=newChartOptions{Plugins=newChartPlugins{Title=newChartTitle{Display=true,Text="Monthly Bar Chart"}},Scales=newScales{Y=newScaleOptions{BeginAtZero=true}}}};ChartRenderer.RenderToPngFile(config,"bar.png",800,500);// or: byte[] png = ChartRenderer.RenderToPng(config, 800, 500);

Render Chart.js JSON directly

Any Chart.js-format configuration deserializes as-is — singular/plural color options, {x,y,r} point data, tension: true, string-or-array colors and numeric paddings all bind the way Chart.js reads them.

usingChartCS.Core;usingChartCS.SkiaSharp;varconfig=ChartJsonOptions.Deserialize("""{ "type": "line", "data": { "labels": ["Sprint 1", "Sprint 2", "Sprint 3"], "datasets": [ { "label": "Completion Rate (%)", "data": [80, 89, 93], "borderColor": "#36a2eb", "backgroundColor": "rgba(54, 162, 235, 0.16)", "borderWidth": 3, "fill": true, "tension": 0.3, "pointRadius": 4 }, { "label": "Commitment Target", "data": [85, 85, 85], "borderColor": "#f67019", "borderWidth": 2, "pointRadius": 0 } ] }, "options": { "plugins": { "title": { "display": true, "text": "Sprint Completion Rate" } }, "scales": { "y": { "beginAtZero": true, "max": 100 } } }}""");ChartRenderer.RenderToPngFile(config,"sprint.png",800,500);

Fluent API

usingChartCS.Core;usingChartCS.Fluent;usingChartCS.SkiaSharp;varconfig=ChartBuilder.Create(ChartType.Line).Title("Weekly Active Users").Legend(position:LegendPosition.Bottom).Labels("Mon","Tue","Wed","Thu","Fri","Sat","Sun").Dataset(d =>d.Label("This week").Data(420,465,490,470,540,380,350).BorderColors("#36a2eb").BackgroundColors("rgba(54, 162, 235, 0.2)").Fill().Tension(0.4).PointRadius(4)).Dataset(d =>d.Label("Last week").Data(390,410,440,460,500,420,390).BorderColors("#ff6384").Tension(0.4).PointRadius(4)).BeginAtZero().Build();ChartRenderer.RenderToPngFile(config,"fluent.png",800,500);

Gallery

Every image above is produced by the demo project — run it yourself:

dotnet run --project samples/ChartCS.Demo # writes PNGs to samples/out/

The demo also renders every *.json file in samples/ChartCS.Demo/samples/ (one Chart.js-format config per chart type) — drop your own config there to try it.

Packages

Package	Contents
ChartCS.Core	Configuration model (datasets, scales, plugins) + Chart.js JSON (de)serialization. No rendering dependencies.
ChartCS.SkiaSharp	The PNG renderer (ChartRenderer.RenderToPng/RenderToPngFile).
ChartCS.Fluent	ChartBuilder/DatasetBuilder chainable configuration API.

Chart.js feature support

Implemented: category & linear scales with Chart.js nice-number ticks, min/max/suggestedMin/suggestedMax/beginAtZero, axis titles & grid & autoskip & reverse, grouped/stacked bars with barPercentage/categoryPercentage, line fills, tension, cubicInterpolationMode: "monotone", spanGaps, showLine, doughnut cutoutPercentage, per-slice arc colors with white borders, legends (per dataset / per slice, position, align, reverse, usePointStyle), titles & subtitles, bubble pixel radii.

Not implemented yet: horizontal bars (indexAxis: "y"), multiple/custom axes, logarithmic & time scales, mixed chart types, dashed borders, point shapes other than circle, tooltips & animation (meaningless for static images). Details in docs/ARCHITECTURE.md.

License

MIT