diff --git a/src/config.js b/src/config.js
index 1a77873dc..399e6fc40 100644
--- a/src/config.js
+++ b/src/config.js
@@ -289,3 +289,26 @@ export const SESSION_MAX_AGE = fromEnvOrConfig("SESSION_MAX_AGE", 30 * 24 * 60 *
 export const GROUPS_DATA_FILE =
      process.env.GROUPS_DATA_FILE
   ?? path.join(__basedir, "env", (PRODUCTION ? "production" : "testing"), "groups.json");
+
+
+/**
+ * Name of the S3 bucket to use for Nextstrain Groups storage.
+ *
+ * The layout of objects in the bucket will be like so:
+ *
+ *   ${bucketName}/
+ *     ${groupName}/
+ *       group-logo.png
+ *       group-overview.md
+ *       datasets/
+ *         ${datasetName}.json
+ *         ${datasetName}_${sidecarType}.json
+ *         …
+ *       narratives/
+ *         ${narrativeName}.md
+ *         …
+ *
+ * @type {string}
+ * @default nextstrain-groups
+ */
+export const GROUPS_BUCKET = fromEnvOrConfig("GROUPS_BUCKET", "nextstrain-groups");
diff --git a/src/sources/groups.js b/src/sources/groups.js
index b85813c63..4823a0dd9 100644
--- a/src/sources/groups.js
+++ b/src/sources/groups.js
@@ -3,6 +3,7 @@ import AWS from 'aws-sdk';
 import { Buffer } from 'buffer';
 import yamlFront from 'yaml-front-matter';
 import * as authz from '../authz/index.js';
+import { GROUPS_BUCKET } from '../config.js';
 import { fetch } from '../fetch.js';
 import { Group } from '../groups.js';
 import * as utils from '../utils/index.js';
@@ -10,7 +11,7 @@ import { Source, Dataset, Narrative } from './models.js';
 
 const S3 = new AWS.S3();
 
-const BUCKET = "nextstrain-groups";
+const BUCKET = GROUPS_BUCKET;
 const DATASET_PREFIX = "datasets/";
 const NARRATIVE_PREFIX = "narratives/";