Stripe built its internal AI platform

On July 30, 2026, Anna Mason, Sharadh Krishnamurthy, and Anupam Upadhyay present Stripe’s Knowledge AI Platform, an internal engineering initiative that centralizes knowledge and tooling to accelerate AI agent development across the company. Built by the Agent Foundation and AI Platform teams, the platform ties together documentation, prototyping tools, and developer workflows so engineers can more quickly build, test, and deploy agent-driven features. Related internal projects referenced include Harbor, an AI-assisted prototyping tool that compiles prototypes in the browser, and WebMCP, which helped speed agent-driven purchases from the browser. The write-up is aimed at developer audiences and links to in-depth docs, community channels like Discord and YouTube, and local meetups to support adoption and collaboration.

Substantively, the platform’s value proposition is operational: it provides a shared knowledge layer and integrated tooling to reduce friction when designing agent experiences - examples include efforts to adapt Checkout for AI agents and streamline agent purchases. The content highlights team ownership, practical resources, and concrete internal tooling rather than speculative research, positioning the platform as a production-ready infrastructure for embedding generative AI into Stripe’s products and developer-facing offerings. The post functions as both an announcement of capability and a pointer to technical guidance and community support for practitioners inside and adjacent to Stripe.