Designing the Layered Tool Pattern
MCP Dev Summit
Shared how the Square MCP server orchestrates discovery, planning, and execution layers so MCP clients like Goose stay reliable even across hundreds of endpoints.
MCP Dev Summit
Shared how the Square MCP server orchestrates discovery, planning, and execution layers so MCP clients like Goose stay reliable even across hundreds of endpoints.
WorkOS MCP Night 2.0
Demoed how the Square MCP server’s layered tool pattern compresses 30+ APIs into discovery, planning, and execution tools that Goose, our MCP client, can reason over.
Square Developer Podcast
Interviewed Eric and Alex from Headpinz on transforming guest experiences and unifying payments across arcade, bowling, and food service.
Square Developer Podcast
Hosted Peter from Ngrok to discuss how programmable traffic and tunneling solve local dev friction and enable secure AI webhook testing.
Changelog & Friends
Changelog invited me to unpack launching the Square Developer Podcast and why MCP servers change how APIs meet AI agents.
Square Developer Podcast
Hosted Apollo’s Alex Hancock and Watson to talk about how GraphQL fits into modern API platforms and AI planning pipelines.
Square Developer Podcast
Interviewed Monstarlab on scaling customer experiences, kiosk flows, and AI-assisted payments across global brands.
APIMatic Podcast
Discussing the benefits of collaboration between internal API consumers and governance committees, and why APIs should be built to last forever.
Square Developer Podcast
Setting the stage for the Square Developer Podcast with Adam Stacoviak from Changelog. Discussing the vision for developer storytelling at Square.
Changelog YouTube
Live conversation with the Changelog crew on building developer media, AGI optimism, and Battle-tested MCP tooling at Block.
Shippo YouTube
A technical deep dive into the Shippo API architecture, covering object models, webhooks, and shipment creation for developers.
WorkOS Blog
WorkOS covered Block’s layered tool approach that shrinks a massive API surface into three MCP primitives for AI enablement.