# Coordination Server **The infrastructure that introduces devices to each other — then gets out of the way.** Coming soon — join the waitlist at for launch notice and early SDK access. ## The problem with connectivity infrastructure Most "serverless" peer-to-peer systems are anything but. They route traffic through central servers that log it, inspect it, and hold it. The server becomes the single point of trust — and the single point of failure. When it's compromised, breached, or compelled, everything it ever touched is exposed. The coordination server is built on a different premise: **a server that has never seen your data cannot leak it.** ## What it does When two devices need to find each other across the internet, they need something in the middle — briefly. The coordination server handles the introduction, relays the handshake, then steps aside so the two devices talk directly. That's the whole job. Anything beyond it is out of scope by design. ## Why that constraint is the product The coordination server holds no application data. It does not inspect what your devices are exchanging. It has no concept of files, messages, channels, or content. It does not log the payload of the connections it sets up. If we were compelled to produce records of what your devices communicated, we could not — because we never had it. This is not a privacy policy. It is an architecture. > "The cheapest way to defeat a 'harvest now, decrypt later' adversary is to make sure the server has nothing worth harvesting." ## A platform, not a feature The coordination server is intentionally application-agnostic. Adding a new application to your product requires no server changes, no new endpoints, no coordination with us — you write client-side code; the infrastructure already supports it. File sharing, encrypted messaging, real-time collaboration, custom peer protocols — the platform was designed so the list never needs to end. ## The connection layer of Web 4.0 Web 4.0 — the [Independent Internet](https://ap0110.com/web4) — is about owning the connectivity itself, not just your data. That requires a way for devices to find each other without handing the connection to a provider who can log it, inspect it, or cut it. The coordination server is that introduction layer: open, self-hostable, and built so no central operator — us included — sits in the path of what you send. It's how peers meet on a network with no single point to fail or be seized. ## Built in to AP0110 products If you're already building on AP0110, the coordination server is already there. Every AP0110 product uses it by default — the same infrastructure, the same architectural guarantees, without standing anything up. ## Available to everyone else - **Hosted (registered access)** — Use the AP0110-operated coordination server directly in your own products. Register for access and build against the SDK. No infrastructure to run. - **Self-hosted (free)** — Run your own instance for private use. The server is open and self-hostable. Your devices, your infrastructure, no dependency on us. - **Self-hosted (enterprise)** — Production-grade self-hosted deployments for organizations with compliance, data-residency, or scale requirements. Supported, auditable, yours. ## The SDK The coordination server comes with an SDK for building products and services on top of it. Device identity, contact management, connection lifecycle, and the signaling layer are handled for you. The wire interface is stable and documented — everything the server and client agree on is a contract, not an implementation detail. ## What you get - **Devices find each other.** Across NATs, across networks, across the open internet. - **Direct connections.** Once introduced, devices communicate directly — traffic does not route through our servers. - **Cryptographic identity.** Each device generates its own keypair locally; identity is verified by the devices themselves, not the server. - **Opaque signaling.** The server sees envelopes, not contents. Your application protocol is your business. - **Rate limiting at the boundary.** Abuse is caught per device, per IP, before any signal is forwarded. ## What it doesn't do It does not store your data, log your connections, or know which devices trust each other. It has no user accounts, no access control lists, no content policies — because it has no content to enforce them against. If you need a server that mediates ongoing application state or can be subpoenaed for conversation history, this is not it. If you need devices to connect directly, privately, and verifiably — this is exactly it. ## Join the waitlist Coming soon. Join the waitlist for launch notice and early SDK access: