The Matrix Conference 2026

Element Call already proves what is possible with MatrixRTC: decentralised, federated, end-to-end encrypted video and voice calls powered entirely by the Matrix protocol — but most of the Matrix ecosystem hasn't yet adopted it.
To change this we need to build an ecosystem around MatrixRTC that is as vibrant as Matix itself.
This talk outlines the vision of components for such an ecosystem and consolidates all the details required to efficiently build on top of MatrixRTC or add support to clients and sdks.

This talk starts with a quick tour of MatrixRTC from the ground up. We start with the core specification (MSC4143): what a MatrixRTC session actually is, how participants announce themselves via m.call.member sticky events, and how a call across a federated network of homeservers and multiple LiveKit SFUs (Selective Forwarding Unit) works.

With the foundations in place, we explore what MatrixRTC makes possible beyond video calls: low-latency data channels, real-time collaborative features, game state synchronisation, and more — all authenticated by Matrix and federated by default.

The second half of the talk will be a conversation about how to build on MatrixRTC correctly and why it matters. We share our learning how to structure SDKs around MatrixRTC and evaluate possible shared building blocks. The goal is to build an ecosystem with components suited for most tasks without reinventing the wheel.

Attendees will gain a basic understanding of the technology so you leave the room knowing exactly where to start, and an awareness of what components are available to make your life easier and the final products stable and compatible. Let's bring MatrixRTC to the whole Matrix ecosystem.

Have you ever wanted to use Matrix as a backend for privately shared notes, workout tracking, expense management, task lists, or even local LLM memory storage?

Matrix is a surprisingly strong fit for these use cases. It offers end-to-end encryption, solid SDK support, multiple homeserver implementations, and effortless redundant storage — simply join a room with two accounts on different homeservers.

All of this makes Matrix an ideal backend for CRDT data structures, particularly for apps that store personal data with owner-controlled collaboration and sharing. In this talk, I will present what requirements Matrix needs to fulfill to be the perfect container for CRDT data, where it currently falls short, and why the path forward is bright — requiring only minimal changes to the spec.

This talk grew out of discussions at the Matrix Conference 2025. At the Matrix Community Summit 2026, those topics were revisited and iterated across multiple widget and integration workshops.