diff --git a/swh/indexer/metadata.py b/swh/indexer/metadata.py
--- a/swh/indexer/metadata.py
+++ b/swh/indexer/metadata.py
@@ -164,7 +164,7 @@
 
         """
         result = {
-            'id': rev['id'].decode(),
+            'id': rev['id'],
             'indexer_configuration_id': self.tool['id'],
             'translated_metadata': None
         }
@@ -287,12 +287,12 @@
 
         Args:
 
-            * `origin_head` (dict): {str(origin_id): rev_id.encode()}
+            * `origin_head` (dict): {str(origin_id): rev_id}
               keys `origin_id` and `revision_id`, which is the result
               of OriginHeadIndexer.
             * `policy_update`: `'ignore-dups'` or `'update-dups'`
         """
-        origin_head_map = {int(origin_id): rev_id
+        origin_head_map = {int(origin_id): hashutil.hash_to_bytes(rev_id)
                            for (origin_id, rev_id) in origin_head.items()}
 
         # Fix up the argument order. revisions_metadata has to be the
diff --git a/swh/indexer/origin_head.py b/swh/indexer/origin_head.py
--- a/swh/indexer/origin_head.py
+++ b/swh/indexer/origin_head.py
@@ -11,6 +11,8 @@
 from swh.scheduler.utils import create_task_dict
 from swh.indexer.indexer import OriginIndexer
 
+from swh.model.hashutil import hash_to_hex
+
 
 class OriginHeadIndexer(OriginIndexer):
     """Origin-level indexer.
@@ -69,7 +71,7 @@
             'oneshot',
             origin_head={
                 str(result['origin_id']):
-                    result['revision_id'].decode()
+                    hash_to_hex(result['revision_id'])
                 for result in results},
             policy_update='update-dups',
             )
@@ -80,7 +82,7 @@
         task = create_task_dict(
             revision_metadata_task,
             'oneshot',
-            ids=[res['revision_id'].decode() for res in results],
+            ids=[hash_to_hex(res['revision_id']) for res in results],
             policy_update='update-dups',
             next_step=sub_task,
             )
diff --git a/swh/indexer/storage/db.py b/swh/indexer/storage/db.py
--- a/swh/indexer/storage/db.py
+++ b/swh/indexer/storage/db.py
@@ -364,7 +364,7 @@
                           regconfig=regconfig,
                           tsquery_template=tsquery_template)
         cur.execute(query, tsquery_args + [limit])
-        yield from cur
+        yield from cursor_to_bytes(cur)
 
     indexer_configuration_cols = ['id', 'tool_name', 'tool_version',
                                   'tool_configuration']
diff --git a/swh/indexer/tests/test_metadata.py b/swh/indexer/tests/test_metadata.py
--- a/swh/indexer/tests/test_metadata.py
+++ b/swh/indexer/tests/test_metadata.py
@@ -14,6 +14,8 @@
 from swh.indexer.tests.test_utils import MockObjStorage, MockStorage
 from swh.indexer.tests.test_utils import MockIndexerStorage
 
+from swh.model.hashutil import hash_to_bytes
+
 
 class ContentMetadataTestIndexer(ContentMetadataIndexer):
     """Specific Metadata whose configuration is enough to satisfy the
@@ -448,14 +450,14 @@
         metadata_indexer = RevisionMetadataTestIndexer()
 
         sha1_gits = [
-            b'8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f',
+            hash_to_bytes('8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f'),
         ]
         metadata_indexer.run(sha1_gits, 'update-dups')
 
         results = metadata_indexer.idx_storage.added_data
 
         expected_results = [('revision_metadata', True, [{
-            'id': '8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f',
+            'id': hash_to_bytes('8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f'),
             'translated_metadata': {
                 '@context': 'https://doi.org/10.5063/schema/codemeta-2.0',
                 'url':
diff --git a/swh/indexer/tests/test_origin_metadata.py b/swh/indexer/tests/test_origin_metadata.py
--- a/swh/indexer/tests/test_origin_metadata.py
+++ b/swh/indexer/tests/test_origin_metadata.py
@@ -16,6 +16,8 @@
 
 from swh.scheduler.tests.scheduler_testing import SchedulerTestFixture
 
+from swh.model.hashutil import hash_to_bytes
+
 
 class OriginMetadataTestIndexer(OriginMetadataIndexer):
     def prepare(self):
@@ -104,13 +106,14 @@
             'keywords': ['yarn', 'parse', 'lock', 'dependencies'],
         }
         rev_metadata = {
-            'id': '8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f',
+            'id': hash_to_bytes('8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f'),
             'translated_metadata': metadata,
             'indexer_configuration_id': 7,
         }
         origin_metadata = {
             'origin_id': 54974445,
-            'from_revision': '8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f',
+            'from_revision': hash_to_bytes(
+                '8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f'),
             'metadata': metadata,
             'indexer_configuration_id': 7,
         }
diff --git a/swh/indexer/tests/test_utils.py b/swh/indexer/tests/test_utils.py
--- a/swh/indexer/tests/test_utils.py
+++ b/swh/indexer/tests/test_utils.py
@@ -5,6 +5,7 @@
 
 from swh.objstorage.exc import ObjNotFoundError
 from swh.model import hashutil
+from swh.model.hashutil import hash_to_bytes
 
 ORIGINS = [
         {
@@ -119,7 +120,8 @@
         54974445: {
                 'branches': {
                     b'HEAD': {
-                        'target': b'8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f',
+                        'target': hash_to_bytes(
+                            '8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f'),
                         'target_type': 'revision'}}}
         }
 
@@ -320,10 +322,12 @@
         self.added_data.append(
                 ('revision_metadata', conflict_update, metadata))
         for item in metadata:
+            assert isinstance(item['id'], bytes)
             self.revision_metadata.setdefault(item['id'], []).append(item)
 
     def revision_metadata_get(self, ids):
         for id_ in ids:
+            assert isinstance(id_, bytes)
             yield from self.revision_metadata.get(id_)
 
     def origin_intrinsic_metadata_add(self, metadata, conflict_update=None):
@@ -387,7 +391,7 @@
 
     def revision_get(self, revisions):
         return [{
-            'id': b'8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f',
+            'id': hash_to_bytes('8dbb6aeb036e7fd80664eb8bfd1507881af1ba9f'),
             'committer': {
                 'id': 26,
                 'name': b'Andrew Nesbitt',