Stripe Now Takes 93% of Payment SDK Downloads on npm
In the four weeks to 20 September, the official Stripe library for Node.js was downloaded 67.1 million times from npm, the public registry most JavaScript developers install code from. That was 92.9% of all downloads of the nine official gateway server libraries HaiPay tracked. A year earlier Stripe's share was 87.6%.
September 22nd, 2026
Last updated: September 22
Key takeaways
- Stripe's Node.js library took 92.9% of npm downloads of nine official gateway server SDKs in four weeks.
- A year earlier its share was 87.6%; it is now 13 times the other eight combined, up from 7 times.
- Twelve non-payment npm packages grew a median 2.45 times over the same windows, the baseline for comparison.
- Against that baseline Square and Stripe gained ground; Adyen, Braintree and Checkout.com lost it.
- The x402 agent payments package had 1.30 million downloads, more than Adyen's or Braintree's server library.
- Downloads count installs, including automated ones; they are not a measure of merchants or volume.
Data highlight
92.9per cent of downloads
Stripe's share of npm downloads of nine official payment gateway server SDKs
24 August to 20 September 2026
Counted by HaiPay on 22 September 2026 from the npm registry downloads API for 24 August to 20 September 2026 and 25 August to 21 September 2025. Nine Node.js server libraries were included, each only where the npm registry lists the gateway's own repository: stripe, @adyen/api-library, braintree, @paypal/paypal-server-sdk, @paypal/checkout-server-sdk, square, @mollie/api-client, razorpay and checkout-sdk-node; @airwallex/node-sdk was excluded because the registry names no repository for it. Their combined downloads were 17,788,030 in the 2025 window and 72,243,464 in the 2026 window. stripe accounted for 67,094,786 in 2026, 92.9 per cent, against 15,584,915 and 87.6 per cent in 2025, and was 13.0 times the other eight combined against 7.1 times. Growth rates were 4.31 for stripe, 4.80 for square, 3.47 for razorpay, 3.57 for Mollie and PayPal's current server SDK, 1.88 for Checkout.com, 1.71 for Adyen, 1.24 for Braintree and 0.86 for PayPal's deprecated Checkout server SDK. Over the same windows the median growth of twelve widely used non-payment packages, chosen by HaiPay, was 2.45, with eleven of the twelve rising; dividing by that baseline gives 1.96 for square, 1.76 for stripe, 1.46 for Mollie and PayPal's current SDK, 1.41 for razorpay, 0.77 for Checkout.com, 0.70 for Adyen, 0.51 for Braintree and 0.35 for the deprecated PayPal SDK. No package's highest day exceeded twice its median day. The x402 package, whose registry entry lists x402-foundation/x402, had 1,298,647 downloads in the 2026 window against 33,193 in 2025, and @x402/core, first published in December 2025, had 999,624. Downloads count installs including automated ones, cover Node.js server libraries only, and are not a measure of merchants, developers or payment volume. HaiPay did not contact any company named.
n the four weeks to 20 September, the official Stripe library for Node.js was downloaded 67.1 million times from npm, the public registry most JavaScript developers install code from. That was 92.9% of all downloads of the nine official gateway server libraries HaiPay tracked.
A year earlier Stripe's share was 87.6%. It is now thirteen times the other eight combined, up from seven times.
What was measured
npm publishes daily download counts for every package through a public API. HaiPay took the server-side libraries that nine payment companies publish for Node.js, and included a package only where the registry lists the company's own code repository: Stripe, Adyen, Braintree, PayPal's current and deprecated server SDKs, Square, Mollie, Razorpay and Checkout.com. Airwallex's package was left out because the registry does not name a repository for it.
The windows are 24 August to 20 September 2026 and the matching four weeks of 2025, 25 August to 21 September, both starting on a Monday. No single day in any package was more than twice that package's median day, so no one spike drives the totals.
Everything grew, so growth alone means little
The first thing the data shows is that almost every library grew fast. Stripe's downloads rose 4.3 times. Square's rose 4.8 times, Razorpay's 3.5 times, Mollie's and PayPal's current server SDK's 3.6 times each.
That is not all payments. HaiPay ran the same comparison on twelve widely used libraries with nothing to do with payments, from lodash and express to twilio and nodemailer. Their median growth over the same windows was 2.45 times. Lodash, a utility library that has barely changed in years, nearly doubled. Eleven of the twelve rose.
Downloads count installs, including automated installs by build systems, test pipelines and tools, not people. Whatever pushed installs up across npm in the last year, it pushed up payment libraries too. So the useful number is growth measured against that baseline.
Against the baseline, the picture splits
Divided by 2.45, five of the nine grew faster than npm as a whole: Square at 1.96, Stripe at 1.76, Mollie and PayPal's current server SDK at 1.46, and Razorpay at 1.41.
Four grew more slowly. Checkout.com's library came in at 0.77, Adyen's at 0.70 and Braintree's at 0.51. PayPal's older Checkout server SDK, which the registry marks as deprecated, was at 0.35, which is roughly what a library being retired should look like.
In share terms, Adyen went from 3.0% of these downloads to 1.3%, and Braintree from 3.1% to 1.0%. Those are not collapses in absolute downloads; both still grew. They are libraries growing more slowly than the registry they live in, in a market where one library already accounted for nearly nine in ten downloads.

An agent payments library passed two incumbents
One package outside the nine is worth setting beside them.
x402 is an open protocol for payments made by software agents. Its main npm package was downloaded 1.30 million times in the four weeks to 20 September, up from 33,193 in the same weeks of 2025. That is more than Adyen's server library, at 915,030, and more than Braintree's, at 687,946. A companion package, @x402/core, first published in December 2025, added nearly a million more.
That comparison needs care. x402 is a set of libraries for building agent payment flows, not a gateway; a download of it is not a merchant choosing it over Adyen. But in developer attention, the kind npm can measure, an agent payments protocol has moved from a rounding error to ahead of two established gateway libraries within a year.

What these numbers are not
They are not market share. Most large merchants integrate through several channels, many never touch these libraries, and a single business can generate thousands of installs a day from automated builds. They cover Node.js only, not Java, Python, PHP or mobile, and server libraries only, not the browser-side components gateways also publish.
They are also not evidence of why installs rose across npm. HaiPay did not test explanations and does not offer one.
HaiPay did not contact any of the companies named.
What to watch
Whether Adyen and Braintree stay below the npm baseline in the next quarter, which would make this a trend rather than a season.
Whether x402's downloads keep pace now that its core package has been split out, because a combined figure is the fairer comparison from here.
And the baseline itself. If installs across npm keep compounding at this rate, raw download counts for any payment library will say less every month, and the relative measure is the only one worth quoting.
This piece reads public registry statistics. It is not investment advice.
How to cite
HaiPay News, "Stripe Now Takes 93% of Payment SDK Downloads on npm", https://www.haipay.net/news/npm-payment-sdk-downloads-stripe-share, September 22nd, 2026
About the author
Crystal
Digital Public Relations
A digital PR specialist with a Master's in Journalism & Communication from UNSW. Started as an intern at ABC Australia, now leads public relations at Haipay, crafting press releases and media strategies that bring brand stories to life.
Reviewed by WeiJun TangEditorial policy








