Tiles: Own your AI with open models and decentralized protocols

As AI becomes central to how we work and communicate, I believe we’re seeing a new form of technological feudalism, where a handful of platforms control the models, identities, data, and distribution channels people depend on. Local-first AI offers an alternative, but many tools still lack the identity, sync, and collaboration features people expect.

In this talk, I’ll share what I’ve learned building Tiles, a local-first, private AI assistant. I’ll discuss how we combine local models, peer-to-peer encrypted sync, the AT Protocol, and DIDs and UCANs for zero-trust authentication and authorization to deliver privacy without sacrificing the convenience of a centralized app. I’ll also introduce Tilekit, the developer API for building custom clients on top of Tiles.

• Building a user-owned AI stack with decentralized identifiers (DIDs) and zero trust authorization with UCANs for identity and authorization.

• Peer-to-peer encrypted sync of chats, and secure funneling of local endpoints for remote links with Iroh.

• Using AT Protocol to add social, and collaboration features to local-first AI applications.

• Deploying and customizing local AI models with declarative Modelfiles.

• Lessons learned from integrating and polishing local models with the Pi agent harness.