diff --git a/backend/src/main/resources/db/migration/V61__add_document_date_index.sql b/backend/src/main/resources/db/migration/V61__add_document_date_index.sql
new file mode 100644
index 00000000..cba62619
--- /dev/null
+++ b/backend/src/main/resources/db/migration/V61__add_document_date_index.sql
@@ -0,0 +1,9 @@
+-- Index on documents.meta_date for the timeline density aggregation (issue #385).
+-- The new GET /api/documents/density endpoint does GROUP BY date_trunc('month', meta_date)
+-- across the full table; an index keeps the aggregation cheap as the archive grows.
+-- Cheap to add at any size and removes the future-investigation tax.
+--
+-- Note: the entity field is `documentDate` but it's mapped to the PostgreSQL column
+-- `meta_date` (see Document.java @Column(name = "meta_date")).
+
+CREATE INDEX IF NOT EXISTS idx_documents_meta_date ON documents (meta_date);