Remix Jam 2026

Enter from Waterloo Terrace on the west side of 620 King St West. Check in, grab your badge, and head upstairs for breakfast.

In a world where agents write most of the code, we should expect a whole lot more from a web framework. That's the idea behind Remix 3. We set out to make every piece faster, smaller, and easier to use, from the server to the browser. It's built for you and your agents, and you can use as much or as little of it as you want. In this talk Michael will show you what we've been working on and where we're going from here.

Michael Jackson and Brooks Lybrand take questions after the opening talk. Ask them about Remix 3's ideas, its tradeoffs, the decisions behind it, or whatever else you've been waiting to ask the Remix team.

Take a break and enjoy lunch with fellow attendees. This is a great time to discuss the morning talks and make new connections.

What if you could skip the bundler build step and start your production server immediately? Remix ships with an asset server designed for production use, where TypeScript, JSX, CSS and static files compile when requested, without sacrificing the developer experience we've come to expect from modern tooling. In this talk we'll explore how the Remix philosophy changes the way we think about bundling, and why we think it's time to give "unbundling" a try.

By the time an application returns a response, Remix has already handled routing, middleware, rendering, and data—but a full-stack framework shouldn’t stop at the server boundary.

Remix extends up into the browser, where frames provide incremental navigation and granular updates to route-owned regions, while the SPA package runs Remix routes entirely on the client. It also extends out into the test runner, with tools for testing routes directly, exercising components in real browsers, and verifying complete end-to-end flows.

In this talk, we’ll explore how these pieces fit together and how Remix carries a consistent, web-standard programming model across server, browser, and test environments.

When we first announced the direction we were taking Remix, we shared a few principles. The first (and most controversial at the time) was “Model-First Development.” The premise was that Remix should be easy not only for people to understand, but also for large language models to generate, navigate, and modify.

More than a year later, we’ve used coding agents across a wide range of Remix projects: building the Remix Store (with React Router), migrating remix.run from React, creating demos and proofs of concept, publishing newsletter archives, and finally rebuilding the Store with Remix 3. Along the way, we’ve learned what makes Remix especially usable for agents, where they get stuck, and where our patterns and abstractions either help or get in the way.

This talk traces that feedback loop: how Brooks works with coding agents, what those agents reveal about Remix (the good, bad, and ugly), and how those lessons are shaping the framework itself.

Make a quick pit stop for snacks and coffee before the final talks.

What happens when you take the ideas behind Remix and apply them to everything you build?

After months of building real apps with Remix 3, Sergio ended up growing an entire stack around it, from i18n and background jobs to email, MCP, and UI components. In this talk, Sergio will share how that happened, why so much of it starts with the Web Platform, and what he's learned about building the Remix way.

Every Shopify Edition asks the browser to do more. This is the story of how we started asking it to do less.

Once we cross the finish line, stick around for food, drinks, and time with the speakers and fellow attendees.