Stripe logo

Stripe

Lessons and lines of code from Stripe engineers

Try:
Stripe logo
Stripe

Developing an open standard for agentic commerce

The post describes the ACP open standard for agentic commerce, enabling programmatic checkout flows between buyers, AI agents, businesses, and payment providers. It explains that ACP can connect with any commerce backend and payments infrastructure, avoiding custom integrations for each AI agent. It covers the transaction workflow, including saving or adding payment credentials, issuing secure tokens for credential sharing, responding to fraud signals, and establishing the merchant of record. The design supports multiple commerce types (physical, digital, subscriptions) and flows such as multi-merchant carts and asynchronous purchases, with emphasis on security and trust. It provides guidance to get started via the ACP specification and related docs, and notes the open-source Apache 2.0 license and compatibility with any payment provider.

Stripe logo
Stripe

High-growth companies stand out with flexible pricing

Stripe summarizes findings from a survey of 2,000+ businesses showing that high-growth companies use flexible pricing—frequent adjustments, usage-based or hybrid models, and experimentation—to drive growth; the post promotes Stripe’s report and its usage-based billing capabilities.

Stripe logo
Stripe

How we built it: Real-time analytics for Stripe Billing

Stripe rebuilt Billing analytics from a daily-batch system to an end-to-end streaming architecture: subscription and invoice updates are processed into a Flink stateful stream, initial historical state is generated via Spark batch jobs (outputting verifiable flat files), and query-time windowed aggregations run on Apache Pinot v2 to deliver low-latency, consistent Dashboard analytics while supporting user-customizable metric definitions.

Stripe logo
Stripe

A framework for pricing AI products

While businesses are rapidly building AI products, monetization remains a challenge. In this post, we share a framework for building a successful pricing str...

Stripe logo
Stripe

The conversion paradox: 3DS trends in regulated markets

This post examines 3DS (three-domain secure) two-factor authentication trigger rates in regulated markets such as France, the UK, and Japan. It finds that despite high 2FA friction, these regions still achieve high checkout conversion rates, highlighting the paradox that stronger security measures do not necessarily harm user conversion.

Stripe logo
Stripe

The top industries and business models using AI for fraud prevention

A survey of over 4,000 payments leaders found that 47% of businesses now use AI tools to detect and prevent fraud, making it the most prevalent AI use case in payments. Adoption varies significantly across industries and business models. The findings highlight which sectors are leading AI-based fraud prevention and the strategies proving most effective.

Stripe logo
Stripe

How we built it: Jurisdiction resolution for Stripe Tax

Stripe built the Jurisdiction Resolution System (JRS) for Stripe Tax by cleaning and centralizing jurisdiction boundary data into a GIS, precomputing time-aware SPOT polygons offline, and performing fast online matching using hierarchical bounding boxes stored in a balanced R-tree (rebuilt with the Sort-Tile-Recursive algorithm) plus final point-in-polygon checks. The design prioritizes low latency (most states under 10 ms) and accurate, time-aware tax calculations while balancing memory and update costs.

Stripe logo
Stripe

Our top product updates from Sessions 2025

We announced the ability to manage multiple payment providers from within Stripe, two new extensibility primitives, support for turnkey consumer issuing prog...

Stripe logo
Stripe

Radar now protects ACH and SEPA payments

Over the last year we’ve seen a 40% increase in the share of noncard payment volume on Stripe. To help you adapt your fraud prevention strategy to this chang...

Stripe logo
Stripe

How we tested the conversion impact of global payment methods

Stripe describes the methodology for an experiment measuring conversion and revenue impact of offering over 50 payment methods. They randomized which payment methods were withheld per session (using hashed composite identifiers to keep treatments consistent across refreshes), validated sample sizes with a three-phase strategy (power analyses, A/A testing, pilot), and analyzed results at scale using causal forest (decision-tree–style) models. The post also references Stripe’s Optimized Checkout Suite and a no-code A/B testing tool.

Stripe logo
Stripe

Testing the conversion impact of 50+ global payment methods

This post examines data from tests across 50+ global payment methods, showing that dynamically surfacing at least one additional relevant method beyond cards leads to a measurable increase in conversion and revenue. The findings are based on controlled experiments that compare user behavior with and without the extra payment options.