Relay Information

Every Relay fetches the relay's NIP-11 information document and exposes it as observables. Use it to show a relay's name and icon, check which NIPs it supports, and respect its limits.

Reading relay information

relay.information$.subscribe((info) => {
  if (!info) return;
  console.log(info.name, info.software, info.supported_nips);
});

// Name and icon fall back to the URL and favicon until the document loads
relay.name$.subscribe((name) => console.log("Name:", name));
relay.icon$.subscribe((icon) => console.log("Icon:", icon));

For one-off checks, getInformation(), getSupported() and getLimitations() return promises.

const supported = await relay.getSupported();
if (supported?.includes(77)) console.log("Relay supports negentropy");

Checking supported NIPs and limits

relay.supported$.subscribe((nips) => console.log("Supported NIPs:", nips));

relay.limitations$.subscribe((limits) => {
  if (!limits) return;
  console.log("Max subscriptions:", limits.max_subscriptions);
  console.log("Max message size:", limits.max_message_length);
});

Caching information documents

Pass a RelayInformationCache to Relay or RelayPool to reuse documents across relay instances, pools and app restarts. The cache has get(key) and set(key, document, ttl) methods, where ttl is a lifetime in milliseconds, so a Keyv instance works as-is.

import Keyv from "keyv";
import { RelayPool } from "applesauce-relay";

const informationCache = new Keyv();
const pool = new RelayPool({ informationCache, informationCacheMs: 60 * 60 * 1000 });

Documents are cached for 24 hours unless informationCacheMs overrides it. Cache keys are normalized relay URLs. Failed fetches are never stored, and cache read or write errors fall back to the fetched document.

Integration

Relay.negentropy() and sync() read getSupported() and fail early when a relay does not list NIP-77. See Negentropy Sync.

Best Practices

• Share one cache instance between pools so each document is fetched once.
• Use persistent storage for the cache in apps that reconnect to the same relays across sessions.
• Treat a null document as unknown, not as "supports nothing". Many relays don't publish one.
• An existing Relay replays its first result, so create a new instance if you need a refreshed document.