The State of JavaScript Charting in 2026

There is no single best JavaScript charting library in 2026, and anyone who tells you otherwise is selling something. The field is mature and crowded, and the libraries that lead differ depending on whether you weigh raw adoption, developer enthusiasm, bundle weight, rendering engine, framework fit, or license. So instead of ranking, we measured. We pulled the real numbers for 13 widely used libraries and mapped where each one actually sits.

The short answer to "which is the most popular": by npm downloads, the low-level toolkit D3 and the veteran Chart.js sit at the top, with ECharts close behind. By GitHub stars, D3 again. But both counts miss an entire category, the commercial libraries that developers pay for and download from a vendor rather than npm. Every way of counting undercounts something, and this post is about reading past the leaderboard.

Key takeaways

• Every popularity metric undercounts something. D3 tops both npm downloads (59.0M) and GitHub stars (113.2k), but it is a toolkit, not a charting library. Chart.js (54.0M) is the most-used library people install on purpose. And npm and stars barely register the commercial libraries (FusionCharts, Highcharts, amCharts), which are sold by license, not install count.
• The field splits three ways, and that split matters more than any ranking: rendering engine (SVG vs canvas vs WebGL), framework model (React-only vs agnostic), and license (open source vs commercial).
• Bundle size spans about a 14x range, from roughly 49 KB (@visx/xychart) to 680 KB (FusionCharts, full build), and the raw numbers mislead: several of the biggest are also the most tree-shakeable.
• "Free" has conditions, and they are not uniform. Chart.js, D3, ECharts, Plotly, and most of the field are unconditionally free (MIT or Apache). FusionCharts and Highcharts are paid for commercial use, amCharts 5 is free only with an attribution link, and ApexCharts is free for non-commercial use and for companies under $2M in revenue. Check the terms before you ship.
• There is no wrong-by-default choice left. Every library here is viable. The question is fit, and the data below tells you which axis to decide on.

How we measured this

Every number in this post is real, dated, and reproducible. No estimates, no vendor claims.

• npm downloads: the npm downloads API, point/last-month, covering 6 June to 5 July 2026.
• GitHub stars: the GitHub REST API stargazers_count, read on 7 July 2026.
• Bundle size: bundlephobia's size API, the minified and gzipped size of each package's main export, on 7 July 2026.

Three honest caveats you must keep in mind while reading:

1. npm downloads are not developer choice. They count continuous-integration runs, registry mirrors, and transitive installs, so a library bundled inside other popular packages racks up downloads whether or not anyone chose it directly. They also barely see the commercial libraries that ship by direct download and paid license rather than npm.
2. GitHub stars flatter open source and toolkits. Commercial products whose core is closed source look tiny: the amCharts 5 and FusionCharts npm repos are dist mirrors, so their star counts (440 and under 100) say nothing about real usage.
3. Bundle size is the full build. Modular libraries import much smaller in practice. Treat the bundle figures as an upper bound, not a fixed cost.

We also excluded Chartist. It was popular years ago, but at roughly 198k monthly downloads and no meaningful recent development, it is a legacy option in 2026, not a live contender.

The data, live

The chart below is rendered with ApexCharts (we use our own library for our own research). Toggle between the three metrics. The ApexCharts bar is tinted only so you can find it; the order is the raw data.

Read it metric by metric and a pattern appears. The download chart is topped by a toolkit and the long-lived generalists. The stars chart keeps that same toolkit on top but reshuffles everything below it. The bundle chart inverts the usual assumption that the most capable libraries are the leanest. The rest of this post works through what each of those tells you.

Which JavaScript charting library is the most popular in 2026?

By direct, intentional installs, Chart.js is the most popular dedicated charting library, at 54.0M downloads a month and 67.5k stars. It renders to canvas, it is framework-agnostic, and it has been the default "just give me a chart" choice for years.

D3's position is the one most people misread. Its 59.0M downloads and 113.2k stars lead this list, but D3 is not a charting library. It is a low-level toolkit for binding data to the DOM, and it is the substrate that several libraries here (Victory, billboard.js, visx) build on. Its numbers reflect its foundational status, not that people reach for it to draw a bar chart.

The download figures also hide an entire category. FusionCharts, a long-established commercial suite, shows only about 141k monthly npm installs, but npm is not how most of its users get it: commercial libraries are downloaded from a vendor account under a paid license, and Highcharts and amCharts are undercounted the same way to a lesser degree. Read npm as a measure of open-source, install-from-registry adoption, not of total real-world usage.

So the honest ranking of libraries you would actually install to make a chart, by open-source download volume, is Chart.js, then ECharts (14.7M), then Highcharts (9.7M), then ApexCharts (7.8M). D3 sits above all of them by raw volume, but as a toolkit it belongs in a different category.

Which libraries have the most momentum?

GitHub stars are a rough proxy for developer enthusiasm, and they tell a different story than downloads. After D3 (113.2k), Chart.js (67.5k) and ECharts (66.8k) are nearly tied, which is notable because ECharts is far less discussed in Western developer circles than its star count warrants. It is a genuinely major library, backed by the Apache Foundation, and strongest in dense, analytical, and geospatial visualizations.

Below that top group, visx (20.9k), Plotly.js (18.2k), and Lightweight Charts (16.5k) form a middle tier, followed by a cluster within a few thousand stars of each other: ApexCharts (15.1k), Nivo (14.0k), Highcharts (12.5k), and Victory (11.2k). Note how low Highcharts sits despite being one of the most widely deployed libraries in industry: it is commercial, and commercial users tend not to star a repo. In other words, once you are past the top three, "most loved" is a close race, and star count alone should not decide it for you.

How big are these libraries?

Bundle size spans roughly 49 KB to 680 KB gzipped across this set, but the ranking is the opposite of what most developers assume. The heaviest full build by a wide margin is FusionCharts (about 680 KB), a large commercial suite that ships as one big bundle. Among the open-source options the biggest are the capable, batteries-included libraries: ECharts (359 KB), ApexCharts (164 KB), and Nivo (143 KB) ship a lot of chart types and interaction out of the box. The lightest are either narrowly scoped or aggressively modular: @visx/xychart (49 KB) is one entry point of a component toolkit, and Lightweight Charts (60 KB) does one thing (financial time series) extremely well.

The trap is comparing full builds as if they were fixed costs. ECharts, Chart.js, D3, and visx are all modular: import three chart types and you ship a fraction of the headline number. Batteries-included libraries like ApexCharts and Plotly trade some of that for a simpler mental model, where most features are present without opt-in wiring. Neither approach is wrong. If bundle weight is your hard constraint, favor a modular library and measure your real import; if developer time is the constraint, the larger all-in-one libraries often win.

For the full picture on why the rendering engine drives both size and capability, see our companion piece on SVG vs canvas for charts.

SVG, canvas, or WebGL: what is each library rendering with?

The rendering engine is the single most useful way to categorize a charting library, because it predicts the library's ceiling on data volume and its floor on interactivity and accessibility. Here is where each one sits.

Library	Engine	Consequence
ApexCharts	SVG	Crisp, per-element interaction and accessibility; SVG data-volume ceiling
Highcharts	SVG	Same SVG profile; long-established rendering
FusionCharts	SVG	Long-established commercial SVG suite
Victory	SVG	SVG, also works in React Native
@visx/xychart	SVG	You compose SVG primitives yourself
billboard.js	SVG	D3-based SVG rendering
D3.js	SVG (canvas capable)	You choose; SVG by default
Chart.js	Canvas	Flat cost at high point counts; manual hit-testing
Lightweight Charts	Canvas	Built for large financial series
amCharts 5	Canvas	Canvas-first rendering
Nivo	SVG and Canvas	Per-chart choice of engine
ECharts	Canvas (SVG optional)	Canvas by default, SVG renderer available
Plotly.js	SVG and WebGL	WebGL trace types for very large scatter and 3D

The takeaway: SVG libraries win on interactivity, crisp export, and out-of-the-box accessibility, and hit a ceiling in the low thousands of individual marks. Canvas and WebGL libraries win at high data density and pay for it in inspectability. If accessibility is a requirement, note that SVG is a much better starting point, a finding we measured directly in our charting accessibility audit.

React-specific or framework-agnostic?

Three of the libraries here (Nivo, Victory, visx) are React-only, built as native components for teams that live in React. They are real options with a genuine ergonomic edge inside a React codebase, but none of them dominates the field the way the generalists do.

The bulk of usage still goes to framework-agnostic libraries. Chart.js, ECharts, Highcharts, ApexCharts, Plotly, amCharts, and FusionCharts render anywhere and ship official or community wrappers for React, Vue, and Angular, so you are not locked to one framework. This matters more than it used to: teams increasingly run more than one framework, and a chart library that outlives a framework migration is worth something. An agnostic core plus a thin wrapper keeps its reach in a way a React-only library cannot.

Which of these are actually open source?

This is where developers get caught, so be precise about it. A large part of the field is genuinely open source under permissive licenses: Chart.js, D3, Plotly.js, Nivo, Victory, billboard.js, and visx are MIT or similar, and ECharts and Lightweight Charts are Apache 2.0. You can use all of these commercially at no cost, with no conditions.

Four libraries are not free-for-all open source, and the terms differ enough that you should not lump them together:

• ApexCharts is published under a custom license (the source is public, but it is not MIT). It is free for non-commercial use and for companies under $2M in annual revenue; above that threshold, commercial use needs a paid license. In practice most teams use it for free, but you should not assume it is unconditionally open source. Full disclosure: this is our library, and we would rather state our own terms plainly than have you discover them later.
• FusionCharts is a commercial product with no free production tier: the unlicensed build renders a trial watermark, and shipping without it requires a paid license.
• Highcharts is free for personal and non-commercial projects only. Any commercial use requires a paid license, regardless of company size. "Free" is the wrong word for it in a business context.
• amCharts 5 is free to use only if you keep an attribution link on the chart. Removing the branding requires a commercial license.

If you are choosing a library for a commercial product and cost or licensing terms matter, this axis can decide the question on its own, independent of any technical merit. Check the license and, where there is a revenue threshold, check which side of it you are on.

The full comparison

Everything in one place. Downloads are last 30 days, stars and bundle size as of 7 July 2026. Bundle is the full gzipped build; "n/a" means bundlephobia could not size it (Plotly.js is over 1MB).

Library	Downloads / mo	Stars	Bundle (gzip)	Engine	Framework	License
D3.js	59.0M	113.2k	90 KB	SVG	Agnostic toolkit	Open source (ISC)
Chart.js	54.0M	67.5k	67 KB	Canvas	Agnostic	Open source (MIT)
ECharts	14.7M	66.8k	359 KB	Canvas	Agnostic	Open source (Apache 2.0)
Highcharts	9.7M	12.5k	100 KB	SVG	Agnostic	Commercial (free non-commercial only)
ApexCharts	7.8M	15.1k	164 KB	SVG	Agnostic	Custom (free under $2M revenue)
Nivo	6.4M	14.0k	143 KB	SVG / Canvas	React	Open source (MIT)
Lightweight Charts	3.1M	16.5k	60 KB	Canvas	Agnostic	Open source (Apache 2.0)
Plotly.js	3.1M	18.2k	n/a	SVG / WebGL	Agnostic	Open source (MIT)
@visx/xychart	1.6M	20.9k	49 KB	SVG	React	Open source (MIT)
Victory	1.6M	11.2k	105 KB	SVG	React / RN	Open source (MIT)
amCharts 5	1.2M	0.4k	88 KB	Canvas	Agnostic	Commercial (free with attribution)
billboard.js	0.2M	6.0k	n/a	SVG	Agnostic	Open source (MIT)
FusionCharts	0.1M	0.1k	683 KB	SVG	Agnostic	Commercial (paid)

How to choose a charting library in 2026

Do not start from a popularity ranking. Start from your three real constraints, in this order.

Your situation	Reach for
React app, want native-feeling components	Nivo, Victory, or visx, or a framework-agnostic library with a React wrapper
Any framework, or a multi-framework team	Chart.js, ECharts, ApexCharts, Highcharts, or Plotly with wrappers
Huge datasets or 3D	Plotly.js (WebGL) or a canvas library like ECharts
Financial time series specifically	Lightweight Charts, or ApexCharts / Highcharts for richer annotation
Hard bundle-size budget	A modular library (visx, Chart.js, ECharts) and measure your real import
Accessibility is a requirement	An SVG library with built-in semantics; see our audit
You need full low-level control	D3, accepting that you build the chart yourself
Commercial product, licensing matters	Confirm the license first; FusionCharts, Highcharts, and amCharts are not free for all uses

The honest meta-point: for the data volumes most dashboards actually show (hundreds to low thousands of points), almost every library here performs fine, and the decision comes down to framework fit, license, and how much you want to build yourself versus get for free.

Where ApexCharts fits

Since this is our blog, here is our position stated as a position, not a verdict. ApexCharts is a framework-agnostic, SVG library with official React, Vue, and Angular wrappers, published under a custom license that is free for non-commercial use and for companies under $2M in annual revenue, with a commercial license above that. In this data it sits mid-field: 7.8M downloads and 15.1k stars put it among the established generalists, and its 164 KB full build reflects a batteries-included design, more chart types and interaction without opt-in wiring, rather than a minimal core.

That makes it a strong fit when you want a broad chart library that works across frameworks, renders crisp and accessible SVG, and is free for most teams to use, and it makes it the wrong choice when you need to render tens of thousands of points (reach for a canvas or WebGL library) or when you are shaving every last kilobyte (reach for a modular one). We think that is a fair place to be, and we would rather tell you the tradeoff than pretend it away.

Summary

JavaScript charting in 2026 is a mature field with no single winner. Chart.js leads deliberate adoption of an actual charting library, D3 tops the raw counts as the toolkit others build on, ECharts is a heavyweight underrated in Western developer conversation, and the commercial libraries (FusionCharts, Highcharts, amCharts) barely register on npm because that is not how they are sold. The decision that matters is not the ranking but the three axes underneath it: the rendering engine sets your data-volume ceiling and your accessibility floor, the framework model decides how native the charts feel and how portable they are, and the license can quietly rule a library out of a commercial product. Measure your real constraints against those three, and the right library for your case falls out of the data. We will refresh these numbers next year.

Methodology: npm point/last-month (6 June to 5 July 2026), GitHub stargazers_count and bundlephobia gzip sizes read 7 July 2026. Numbers change; treat them as a snapshot, not a constant.

Frequently asked questions