The Model Context Protocol (MCP) is getting an upgrade!

Join the global community as we celebrate the latest MCP release with an evening of technical talks, community discussions, and networking. Whether you're already building with MCP or just starting to explore the ecosystem, this is an opportunity to learn what's new, meet other builders, and connect with the people shaping the future of AI interoperability.

This event is part of a series of MCP Release Parties taking place simultaneously in cities around the world.

Capacity is limited, so please only register if you plan to attend.

What We'll Cover

We'll explore what's new in the latest MCP release, hear from builders and contributors across the ecosystem, and discuss how the protocol is being adopted in real-world AI applications.

Expect technical talks, community conversations, demos, and plenty of opportunities to connect with fellow developers.

Agenda

17:30 - 18:30 | Doors Open, Catering & Networking

Check in, grab food and drinks, and meet other engineers, founders, and builders working on AI systems.

18:30 - 18:40 | Welcome & Kickoff

Opening remarks from Agentic AI Foundation (AAIF) Community London.

18:40 - 18:55 | MCP Release Overview

In this talk, David will share the MCP 2026-07-28 specification, the reasoning behind it, and what's coming next for MCP. Have your questions ready for a quick Q&A at the end of it.

🎤 David Soria Parra, Member of Technical Staff (co-creator of MCP) @ Anthropic

David Soria Parra is a Member of Technical Staff at Anthropic based in London. He co-created the Model Context Protocol (MCP). He leads its technical direction, governance, and ecosystem development. Before joining Anthropic in 2024, David spent nearly a decade at Meta leading engineering organisations, and served as an Entrepreneur in Residence at Sutter Hill Ventures. He is a long-time open-source contributor with a track record spanning PHP, Git, and Mercurial.

18:55 - 19:15 | Enterprise Managed Auth: Making Your MCP Server Enterprise-Ready

MCP makes it easy for AI models to reach the data and tools people work with every day. But when MCP is deployed across a company, two challenges appear quickly: employees need to connect clients like Claude to each tool manually, and IT teams need visibility and control over which tools are connected, who authorised them, and how access is governed.

Enterprise Managed Auth, built on the IETF Cross-App Access specification, addresses this by placing the enterprise identity provider in the middle of MCP connections.

In this talk, Aaron Parecki will walk through how Enterprise Managed Auth works end to end: from an IT admin setting up Claude in an identity provider and configuring which enterprise apps it can access, to an employee opening Claude and finding everything already connected.

If you are building an MCP server for enterprise customers, this talk will explain the integration path that makes your server IT-admin-friendly from day one.

🎤 Aaron Parecki, Director of Identity Standards @ Okta

Aaron Parecki is the Director of Identity Standards at Okta. He has spent over a decade editing OAuth specifications at the IETF, including several of the specs that are referenced by the MCP authorization profile. Most recently, he created the Cross-App Access (XAA) specification that underlies Enterprise Managed Auth, and has been driving its adoption across the MCP ecosystem.

19:15 - 19:35 | Connecting Context: Stateless MCP

MCP becoming stateless is one of the largest changes to the protocol since its launch.

In this talk, Shaun Smith will explain the motivation for going stateless, sharing data and experience from building popular clients and servers in production.

He will also discuss what builders need to know: what is new, what has changed, and what new use cases this unlocks.

🎤 Shaun Smith, MCP / Open Source @ Hugging Face

Shaun Smith works on MCP and Agents at Hugging Face, and is an active MCP Maintainer, contributor and enthusiast.

19:35 - 19:45 | Break

Refuel, connect, and continue the conversation. There will also be a small MCP-themed surprise during the break…

19:45 - 20:05 | Beyond the Core: MCP Extensions

The latest MCP protocol version introduces a formalised extensions framework - a structured approach to extending MCP beyond its core capabilities. But what makes something "core" versus an extension? How do extensions interact with capability negotiation, and what does adoption look like when both client and server must opt in?

This talk unpacks the extensions framework, surveys the experimental extensions emerging from MCP working groups, and closes with a hands-on look at MCP Apps - the first official extension - exploring not just what they enable but the practical security considerations that come with giving agents a user-facing surface.

🎤 Peder Holdgaard Pedersen, Head of AI Engineering @ Saxo Bank

Peder builds and architects AI systems and spearheads AI adoption at Saxo Bank as Head of AI Engineering. He is a contributor to the C# MCP SDK, an MCP maintainer, and a Microsoft .NET MVP. He specializes in developing bespoke assistants and agents, with focus on safety, security, and compliance at scale.

20:05 - 20:10 | What’s New in the MCP Registry: More Ecosystems, Safer Publishing, Stable APIs

A 5-minute tour of what’s new in the MCP Registry: a stable versioned API (v0.1), server lifecycle management (deprecate/delete with messages), more package ecosystems including crates.io, multi-tenant remote servers, and hardened publishing. Where we are and where we’re headed.

🎤 Radoslav Dimitrov, Staff Software Engineer @ Stacklok

Radoslav Dimitrov is a Staff Software Engineer at Stacklok focused on making AI and MCP easy to use and more secure. He helps maintain the official Model Context Protocol registry and ToolHive and spent a few years working on supply chain security before that

20:10 - 20:40 | Panel Discussion + Audience Q&A

We’ll bring the speakers together for a hosted panel discussion on where MCP is heading, what the latest release unlocks, and what builders should watch as the ecosystem matures.

We’ll then open the discussion up to audience questions.

Bring your questions.

20:40 - 21:00 | Networking

Continue the conversation with speakers, attendees, and other builders working on MCP, agents, and production AI systems.

21:00 - Onwards | Social

Optional post-event social nearby.

Location: shared at the event.

Who Should Come

AI engineers, developers, founders, researchers, open source contributors, and anyone interested in the future of AI interoperability.

Whether you're already building MCP servers and clients or simply curious about the protocol, you'll find people to learn from and connect with.

About AAIF

The Agentic AI Foundation (AAIF) is a community dedicated to advancing the understanding and practical application of agentic AI. We bring together engineers, researchers, founders, builders, and AI enthusiasts to explore how autonomous AI systems are designed, evaluated, and deployed.

Through reading groups, workshops, and community discussions, AAIF brings together engineers, researchers, founders, and AI enthusiasts to exchange ideas, challenge assumptions, and learn from one another.

By attending, you agree to our Code of Conduct and Privacy Policy.

MCP Release Party