Keep a local mirror in sync

Want your own database to always reflect what's in a Dib home? You've got two ways to stay current plus a way to catch up after downtime. Pick based on how fresh you need the data.

Option 1: Poll the change feed

The change feed is an append-only log of everything that happens in a team. Each event has a monotonically increasing id; you track a cursor and ask for everything after it. This is the simplest reliable approach and it survives restarts.

// Resume from the cursor you stored last run (or omit for a full backfill).
let cursor = await loadCursor(); // your storage

while (true) {
  const url = new URL("https://dib.io/api/v1/events");
  if (cursor) url.searchParams.set("cursor", cursor);
  url.searchParams.set("limit", "100");

  const res = await fetch(url, {
    headers: { Authorization: `Bearer ${process.env.DIB_API_KEY}` },
  });
  const { data, meta } = await res.json();

  for (const event of data) {
    // event.type is "created" | "updated" | "deleted"
    // event.payload is the post-image of the resource (null on delete)
    await applyToMirror(event);
  }

  if (meta.page.next_cursor) {
    cursor = meta.page.next_cursor;
    await saveCursor(cursor); // persist so a crash resumes cleanly
  }
  if (!meta.page.has_more) break; // caught up — sleep, then poll again
}

Each event looks like this:

{
  "data": [
    {
      "id": 1042,
      "type": "updated",
      "resource_type": "inventory",
      "resource_id": "6f1c2e2a-1b3c-4d5e-8f90-1a2b3c4d5e6f",
      "api_version": "2026-06-01",
      "payload": { "id": "6f1c...", "name": "Bosch Dishwasher", "room_id": "..." },
      "created_at": "2026-06-17T22:14:03.000Z"
    }
  ],
  "meta": {
    "request_id": "req_abc123",
    "page": { "has_more": true, "next_cursor": "eyJ2IjoxLCJpZCI6MTA0Mn0" }
  }
}

A few things worth knowing: type is created, updated, or deleted; payload is the full post-image of the resource (and is null for deletes, so lean on resource_id there). Store next_cursor after each batch so a crash picks up exactly where it left off. Filter to what you care about with ?types=inventory,task.

Option 2: Stream over SSE

Need changes the instant they happen? Open a Server-Sent Events connection and Dib pushes events as they land, with a 60-second heartbeat. Reconnect with your last cursor and you won't miss a thing.

curl -N -H "Authorization: Bearer $DIB_API_KEY" \
     "https://dib.io/api/v1/events/stream"

Catching up after downtime

If your consumer was offline longer than the feed's retention, do a reconciliation pass per collection using updated_since with sort=updated_at. Page through with the cursor until has_more is false, then resume the live feed.

# Pull everything changed since your last sync, oldest-first, per resource.
curl -H "Authorization: Bearer $DIB_API_KEY" \
     "https://dib.io/api/v1/inventory?updated_since=2026-06-01T00:00:00Z&sort=updated_at&order=asc&limit=100"

Next steps

• See every list parameter on the API reference.
• New here? Start with Getting started.