diff --git a/modules/release-notes/pages/relnotes.adoc b/modules/release-notes/pages/relnotes.adoc
index cbf276b5f4..17b545240f 100644
--- a/modules/release-notes/pages/relnotes.adoc
+++ b/modules/release-notes/pages/relnotes.adoc
@@ -6,6 +6,8 @@
 
 include::partial$MB-67762-warning.adoc[]
 
+include::partial$docs-server-7.6.13-release-notes.adoc[]
+
 include::partial$docs-server-7.6.12-release-notes.adoc[]
 
 include::partial$docs-server-7.6.11-release-notes.adoc[]
diff --git a/modules/release-notes/partials/docs-server-7.6.13-release-notes.adoc b/modules/release-notes/partials/docs-server-7.6.13-release-notes.adoc
new file mode 100644
index 0000000000..edbe9a04e7
--- /dev/null
+++ b/modules/release-notes/partials/docs-server-7.6.13-release-notes.adoc
@@ -0,0 +1,51 @@
+[#release-7-6-13]
+== Release 7.6.13 (September 2026)
+Couchbase Server 7.6.13 was released in September 2026.
+This maintenance release contains fixes to issues.
+
+== Fixed Issues
+{empty}
+
+=== Cluster Manager
+
+[#table-7-6-13-Cluster-Manager, cols="10,40,40"]
+|===
+|Issue |Description |Resolution
+
+|https://jira.issues.couchbase.com/browse/MB-65348[MB-65348]
+
+|The Prometheus scrape endpoint lacks a Content-Type header.
+This improvement adds the PrometheusText0.0.4 header to the metrics endpoint.
+It ensures compatibility with monitoring tools.
+|The system now includes the correct Content-Type header for Prometheus metrics.
+
+|===
+
+=== Data Service
+
+[#table-7-6-13-Data-Service, cols="10,40,40"]
+|===
+
+|Issue |Description |Resolution
+
+|https://jira.issues.couchbase.com/browse/MB-74199[MB-74199]
+|DCP sends mutations with an incorrect Snappy datatype bit.
+This occurs after decompressing `xattr`-only mutations.
+The stale bit causes `gocbcore` to drop messages as corrupt.
+|The fix removes the Snappy datatype bit when the emitted payload is uncompressed.
+
+|===
+=== XDCR
+
+[#table-7-6-13-XDCR, cols="10,40,40"]
+
+|===
+|Issue |Description |Resolution
+
+|https://jira.issues.couchbase.com/browse/MB-72304[MB-72304]
+|Rebalanced-out nodes use stale manifests if buckets are re-created.
+This causes issues with document back-filling in XDCR.
+Stale data persists when nodes return to the cluster.
+|The update adds a cleanup process for stale manifest objects on the node store.
+
+|===
\ No newline at end of file