diff --git a/scripts/managed_repo_symlinks.py b/scripts/managed_repo_symlinks.py
new file mode 100644
index 00000000..19fb0140
--- /dev/null
+++ b/scripts/managed_repo_symlinks.py
@@ -0,0 +1,140 @@
+import argparse
+import csv
+import sys
+import os
+import shutil
+
+import omero.clients
+from omero.cli import cli_login
+from omero.gateway import BlitzGateway
+
+
+def create_symlinks(conn, fileset_id, fileset_dirs, args):
+    """ fileset_dirs is dict of 'filset_dir.zarr': 's3_dir_name.zarr' """
+
+    fileset = conn.getQueryService().get("Fileset", fileset_id, conn.SERVICE_OPTS)
+    template_path = os.path.join(args.repo, fileset.templatePrefix.val)
+
+    if args.report:
+        print("\nFileset:", fileset.id.val, template_path)
+    # /data/OMERO/ManagedRepository/demo_2/Blitz-0-Ice.ThreadPool.Server-8/2023-04/07/13-29-24.048/
+
+    preview_image(conn, fileset_id, args)
+
+    # find files/dirs in Fileset template_path that are also in the symlink dir...
+    fs_contents = os.listdir(template_path)
+    if args.report:
+        print("fs_contents", fs_contents)
+
+    for fs_item in fs_contents:
+        # check if item is in target dir
+        target_dir = fileset_dirs.get(fs_item, fs_item)
+        symlink_target = os.path.join(args.target, target_dir)
+        if not os.path.exists(symlink_target):
+            print("Symlink target not found:", symlink_target)
+            continue
+        symlink_source = os.path.join(template_path, fs_item)
+        if args.report:
+            print(f"Link from {symlink_source} to {symlink_target}")
+        if not args.dry_run:
+            # delete directory
+            try:
+                shutil.rmtree(symlink_source)
+            except:
+                # in case the source is already a symlink (or a file)
+                os.remove(symlink_source)
+            # replace with symlink
+            target_is_directory = os.path.isdir(symlink_target)
+            os.symlink(symlink_target, symlink_source, target_is_directory)
+
+
+def preview_image(conn, fileset_id, args):
+
+    params = omero.sys.ParametersI()
+    params.addId(fileset_id)
+    params.page(0, 1)
+    query = "select img from Image img where img.fileset.id=:id"
+    first_image = conn.getQueryService().findAllByQuery(query, params, conn.SERVICE_OPTS)[0]
+
+    if args.report:
+        print("Render Image", first_image.id.val)
+    image = conn.getObject("Image", first_image.id.val)
+    image.renderJpeg()
+
+
+def get_object(conn, obj_string):
+    for dtype in ["Screen", "Plate", "Image", "Fileset", "Dataset"]:
+        if obj_string.startswith(dtype):
+            obj_id = int(obj_string.replace(dtype + ":", ""))
+            obj = conn.getObject(dtype, obj_id)
+            if obj is None:
+                print(obj_string, "not found!")
+            return obj
+ 
+
+def get_fileset_id(conn, obj_string):
+    """obj_string is Image:123 or Fileset:123 or Plate:123"""
+
+    # We never do conn.getObject("Fileset") as it loads
+    # all original files and images - can be too slow
+    if obj_string.startswith("Fileset:"):
+        return int(obj_string.replace("Fileset:", ""))
+
+    obj = get_object(conn, obj_string)
+    if obj_string.startswith("Image:"):
+        return obj.fileset.id.val
+    if obj_string.startswith("Plate:"):
+        well = list(obj.listChildren())[0]
+        image = list(well.listChildren())[0].getImage()
+        return image.fileset.id.val
+
+
+def main(argv):
+    """
+    We go through all Filesets under the top-level Object, and check for items in the
+    templatePrefix dir under ManagedRepository. If these items are also found in
+    the target directory, we delete the item from the Managed Repo and replace it with
+    a symlink to the equivalent dir (or file) in the target dir.
+    If fileset-mappings csv file is given, we map from fileset name e.g. plate1.zarr to a differently
+    named directory within the `target`, e.g. abc123/abc123.zarr (pattern used by BioStudies).
+    Each row of csv should be e.g. "plate1.zarr, abc123/abc123.zarr"
+    """
+    parser = argparse.ArgumentParser()
+    parser.add_argument('object', help='Object:ID where Object is Screen, Plate, Dataset, Image, Fileset')
+    parser.add_argument("target", help="path to dir that contains symlink targets")
+    parser.add_argument("--fileset-mappings", help="Optional path/to/file.csv where each row is: fs_name, symlink_target")
+    parser.add_argument("--repo", help="Managed Repo absolute path", default="/data/OMERO/ManagedRepository")
+    parser.add_argument("--report", action="store_true", help="Print logs")
+    parser.add_argument("--dry-run", action="store_true", help="Don't save any changes")
+    args = parser.parse_args(argv)
+    object_str = args.object
+
+    fileset_dirs = {}
+    # handle fileset_mappings
+    if args.fileset_mappings:
+        with open(args.fileset_mappings, newline='') as csvfile:
+            csvreader = csv.reader(csvfile, delimiter=',')
+            for row in csvreader:
+                fileset_dirs[row[0]] = row[1]
+    if args.report:
+        print("fileset_dirs", fileset_dirs)
+
+    with cli_login() as cli:
+        conn = BlitzGateway(client_obj=cli._client)
+        assert ":" in object_str
+
+        obj_strings = [object_str]
+        if "Screen" in object_str:
+            screen = get_object(conn, object_str)
+            obj_strings = [f"Plate:{plate.id}" for plate in screen.listChildren()]
+        elif "Dataset" in object_str:
+            dataset = get_object(conn, object_str)
+            obj_strings = [f"Image:{image.id}" for image in dataset.listChildren()]
+
+        for object_str in obj_strings:
+            fileset_id = get_fileset_id(conn, object_str)
+            create_symlinks(conn, fileset_id, fileset_dirs, args)
+
+
+if __name__ == '__main__':
+    main(sys.argv[1:])