September Letter to Tokenholders

Our first product under Meridian will be agent email hosted on the user’s private network; the proposed design would route messages sent to an address such as agent@mynetwork.ur.network directly to a mail server on that network. Users could operate their agents’ mailboxes on infrastructure they control, creating alternatives to centralized hosted email providers such as Gmail or Proton. We think that over time, agent identity persistence will be driven by email architecture, whereby they can communicate with other agents, humans and services via an email framework. If a business deploys agents for recurring operational work, such as procurement, scheduling or customer enquiries, the objective is to workflows easier, core features would need to be:

1. Let AI agents handle routine business correspondence
2. Give each agent its own inbox, defined permissions and reviewable activity
3. Keeping sensitive stored messages and mailbox within infrastructure under your control

Our broader vision for Meridian is to help developers deploy AI agents faster and spend less time managing infrastructure. Beyond agent identity i.e. email, we plan to add encrypted code storage and private networking so agents can communicate, access code and connect to internal tools through one integrated service. A lightweight companion process, designed to run without administrator privileges, would make Meridian easier to add to existing deployments. For developers, the intended benefits are fewer integrations to build, less networking to configure and lower ongoing maintenance costs, while retaining control over their data and agent access. In this manner, we think that by accelerating the proliferation of agent capability, we can embed our world class proxy network for their agents too.

The Mycelium Of The Private Internet

Our long-term vision is for privacy to become an expected property of the internet. Our mission is simple: People should be able to communicate, build and transact securely without having to understand the infrastructure protecting them. Accessing and leveraging the internet is a human right, and where there may be gatekeepers who try to suppress access or collect tolls, we aim to stay true to these values.

Brien likes using mycelium as a useful image of how such a network could grow: Beneath a forest floor, countless local connections form a vast structure that is largely invisible from above. UR’s ambition is similarly built from the bottom up: individuals contribute connectivity, independent operators develop services, and developers bring applications and users. Each participant has a practical reason to join, while contributing to infrastructure whose reach extends beyond any single participant.

For the user, that infrastructure should quietly do its job: An agent should be able to receive a message at its own address, exchange private code and connect to another service without requiring its owner to configure a complex network. Achieving that simplicity requires considerable work underneath, particularly in routing, reliability and making services reachable across different environments.

The internet’s adoption of encrypted web connections provides a useful precedent, encryption became part of everyday browsing as software made it routine. We see an opportunity to extend that expectation further, towards private identities and communications hosted on infrastructure users control. Much of today’s internet already encrypts information in transit; our ambition is to reduce the additional exposure and dependence that remain around who is communicating, how they connect and who holds their data. This will be the legacy of UR (SN25).