diff --git a/docs.json b/docs.json
index d91fa81..c10ad83 100644
--- a/docs.json
+++ b/docs.json
@@ -147,6 +147,7 @@
               "references/api/utilities",
               "references/api/get-chains",
               "references/api/get-chains-liquidity",
+              "references/api/get-integrators-usage",
               "references/api/get-currencies-v2",
               "references/api/get-token-price",
               "references/api/transactions-index",
diff --git a/references/api/api-keys.mdx b/references/api/api-keys.mdx
index 1d694d4..08cb5fb 100644
--- a/references/api/api-keys.mdx
+++ b/references/api/api-keys.mdx
@@ -26,6 +26,7 @@ The [Relay Dashboard](https://dashboard.relay.link) is also where you observe wh
 - **Custom views** — save any combination of filters, sort order, and columns as a reusable view for your team, each with its own name, icon, and color.
 - **Keys** — create, rename, and revoke keys for your account.
 - **Webhooks** — configure a per-key [webhook endpoint](/references/api/api_guides/webhooks) so transaction status updates are pushed to your backend instead of polling.
+- **Programmatic usage metrics** — the [Get Integrator Usage](/references/api/get-integrators-usage) endpoint returns aggregated request counts grouped by endpoint, status code, and error code across `minutely` / `hourly` / `daily` buckets, so you can wire the same data the Dashboard shows into your own monitoring.
 
 When debugging a specific transaction, paste the `requestId` from your application into the dashboard's search to jump straight to the matching record.
 
diff --git a/references/api/get-integrators-usage.mdx b/references/api/get-integrators-usage.mdx
new file mode 100644
index 0000000..06ab854
--- /dev/null
+++ b/references/api/get-integrators-usage.mdx
@@ -0,0 +1,11 @@
+---
+title: "Get Integrator Usage"
+description: "Returns aggregated request counts grouped by API key, endpoint, status code, and error code"
+openapi: get /integrators/usage
+---
+
+Programmatic counterpart to the Dashboard's request history — returns the same aggregated counts you see under **Requests**, broken down by endpoint, HTTP status, and (for `400` responses) the machine-readable `errorCode`.
+
+The `granularity` query parameter controls the time bucket and the default window: `minutely` (last 24h), `hourly` (last 7d), or `daily` (last 14d). Data older than 14 days is not retained.
+
+<Tip>Use `granularity=minutely` to build a near-real-time uptime view; use `granularity=daily` for longer-range trend analysis. The `errorCode` field on `400` rows lets you distinguish validation errors (e.g. `AMOUNT_TOO_LOW`, `UNSUPPORTED_CHAIN`) from other client errors when calculating a meaningful success rate.</Tip>