diff --git a/swh/provenance/mongo/backend.py b/swh/provenance/mongo/backend.py
--- a/swh/provenance/mongo/backend.py
+++ b/swh/provenance/mongo/backend.py
@@ -140,11 +140,14 @@
         yield from sorted(occurs, key=lambda x: (x.date, x.revision, x.origin, x.path))
 
     def content_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, datetime]:
+        # FIXME, add index in contnet sha1 and ts
+        # FIXME, do the timezone operation in mongo
         return {
             x["sha1"]: datetime.fromtimestamp(x["ts"], timezone.utc)
+            # FIXME try to avoid this loop and return directly in the needed format from mongo
             for x in self.db.content.find(
-                {"sha1": {"$in": list(ids)}, "ts": {"$ne": None}},
-                {"sha1": 1, "ts": 1, "_id": 0},
+                    {"sha1": {"$in": list(ids)}, "ts": {"$ne": None}},
+                    {"sha1": 1, "ts": 1, "_id": 0},
             )
         }
 
@@ -173,6 +176,8 @@
     def directory_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, datetime]:
         return {
             x["sha1"]: datetime.fromtimestamp(x["ts"], timezone.utc)
+            # FIXME try to avoid this loop and return directly in the needed format from mongo
+            # FIXME add ts to index in directory
             for x in self.db.directory.find(
                 {"sha1": {"$in": list(ids)}, "ts": {"$ne": None}},
                 {"sha1": 1, "ts": 1, "_id": 0},
@@ -180,12 +185,17 @@
         }
 
     def entity_get_all(self, entity: EntityType) -> Set[Sha1Git]:
-        return {
-            x["sha1"]
-            for x in self.db.get_collection(entity.value).find(
-                {}, {"sha1": 1, "_id": 0}
-            )
-        }
+        # only for tests
+
+        return set(self.db.get_collection(entity.value).distinct('sha1'))
+
+        # return {
+        #     x["sha1"]
+        #     # FIXME try to avoid this loop and return directly in the needed format from mongo
+        #     for x in self.db.get_collection(entity.value).find(
+        #         {}, {"sha1": 1, "_id": 0}
+        #     )
+        # }
 
     def location_add(self, paths: Iterable[bytes]) -> bool:
         # TODO: implement this methods if path are to be stored in a separate collection
@@ -194,6 +204,7 @@
     def location_get_all(self) -> Set[bytes]:
         contents = self.db.content.find({}, {"revision": 1, "_id": 0, "directory": 1})
         paths: List[Iterable[bytes]] = []
+
         for content in contents:
             paths.extend(value for _, value in content["revision"].items())
             paths.extend(value for _, value in content["directory"].items())
@@ -216,9 +227,22 @@
                 self.db.origin.insert_one({"sha1": sha1, "url": url})
         return True
 
+        # origins = {
+        #     x["sha1"]: x
+        #     for x in self.db.origin.find(
+        #         {"sha1": {"$in": list(urls)}}, {"sha1": 1, "url": 1, "_id": 1}
+        #     )
+        # }
+        # for sha1, url in urls.items():
+        #     if sha1 not in origins:
+        #         # add new origin
+        #         self.db.origin.insert_one({"sha1": sha1, "url": url})
+        # return True
+
     def origin_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, str]:
         return {
             x["sha1"]: x["url"]
+            # FIXME try to avoid this loop and return directly in the needed format from mongo
             for x in self.db.origin.find(
                 {"sha1": {"$in": list(ids)}}, {"sha1": 1, "url": 1, "_id": 0}
             )
@@ -283,6 +307,7 @@
     def relation_add(
         self, relation: RelationType, data: Dict[Sha1Git, Set[RelationData]]
     ) -> bool:
+
         src_relation, *_, dst_relation = relation.value.split("_")
 
         dst_objs = {
diff --git a/swh/provenance/mongo/bootstrap.py b/swh/provenance/mongo/bootstrap.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/bootstrap.py
@@ -0,0 +1,8 @@
+# FIXME, maybe this should not be part of the code
+
+collections = []
+
+indexes = []
+
+def bootstrap_db():
+    pass
diff --git a/swh/provenance/mongo/errors.py b/swh/provenance/mongo/errors.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/errors.py
@@ -0,0 +1,6 @@
+class DBError(Exception):
+    # FIXME, add mongo specific logging
+    pass
+
+class EntiryError(Exception):
+    pass
diff --git a/swh/provenance/mongo/models/cache.py b/swh/provenance/mongo/models/cache.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/models/cache.py
@@ -0,0 +1,5 @@
+class Cache:
+    """
+    A object cache layer
+    An in memory implementation
+    """
diff --git a/swh/provenance/mongo/models/content.py b/swh/provenance/mongo/models/content.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/models/content.py
@@ -0,0 +1,12 @@
+class Content(Entity):
+    collection = 'content'
+    model = {}
+    validate_model = False
+
+
+    def find_first(sef):
+        pass
+
+
+    def find_all(self):
+        pass
diff --git a/swh/provenance/mongo/models/directory.py b/swh/provenance/mongo/models/directory.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/models/directory.py
@@ -0,0 +1,4 @@
+class Directory(Entity):
+    collection = 'directory'
+    model = {}
+    validate_model = False
diff --git a/swh/provenance/mongo/models/entity.py b/swh/provenance/mongo/models/entity.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/models/entity.py
@@ -0,0 +1,60 @@
+from abc import ABCMeta
+
+
+class Entity(ABCMeta):
+    """
+    An object saved in the db
+    """
+
+    collection: str
+    model: dict
+    validate_model: bool
+
+    @staticmethod
+    def factory(entity: str):
+        mapping = {
+            'content': Content
+            'directory': Directory
+        }
+        if entity in mapping:
+            return entity[mapping]
+        raise EntiryError(f"invalid entity type {entity}")
+
+    def __int__(self, data):
+        self.model = self._load_model()
+        self.data = self._set_data(data)
+
+    def _load_model(self):
+        return {}
+
+    def _set_data(self, data):
+        pass
+
+    def _validate(self):
+        self.db.command()  # use json in schema.json
+
+    def save(self):
+        if self.validate_model and self._validate():
+            raise DataError()
+        save()
+
+    def get(self, qry):
+        pass
+
+    def _is_older_in_time(self):
+        return self.data.ts < ts
+
+    def add_if_older(self):
+        if _is_older():
+            self.save()
+
+    def with_excetion_handle(self):
+        pass
+
+
+class EntityList:
+    """
+    List or array of entities and their operations
+    Operate mostly on object cache
+    """
+    pass
diff --git a/swh/provenance/mongo/models/origin.py b/swh/provenance/mongo/models/origin.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/models/origin.py
@@ -0,0 +1,4 @@
+class Origin(Entity):
+    collection = 'origin'
+    model = {}
+    validate_model = False
diff --git a/swh/provenance/mongo/models/revision.py b/swh/provenance/mongo/models/revision.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/models/revision.py
@@ -0,0 +1,4 @@
+class Revision(Entity):
+    collection = 'revision'
+    model = {}
+    validate_model = False
diff --git a/swh/provenance/mongo/schema.json b/swh/provenance/mongo/schema.json
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/schema.json
@@ -0,0 +1,34 @@
+// The json schema for the first version
+
+{
+  "content": {
+    "bsonType": "object",
+    "required": [ "sha1" ],
+    "properties": {
+      "sha1": {
+        "bsonType": "Binary",
+        "description": ""
+      },
+    }
+  },
+  'directory': {
+    "bsonType": "object",
+    "required": [ "sha1" ],
+    "properties": {
+      "sha1": {
+        "bsonType": "Binary",
+        "description": ""
+      },
+    }
+  },
+  'revision': {
+    "bsonType": "object",
+    "required": [ "sha1" ],
+    "properties": {
+      "sha1": {
+        "bsonType": "Binary",
+        "description": ""
+      },
+    }
+  }
+}
diff --git a/swh/provenance/mongo/storage.py b/swh/provenance/mongo/storage.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/storage.py
@@ -0,0 +1,104 @@
+# Copyright (C) 2021  The Software Heritage developers
+# See the AUTHORS file at the top-level directory of this distribution
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+from datetime import datetime, timezone
+import os
+from typing import Any, Dict, Generator, Iterable, List, Optional, Set, Union
+
+from bson import ObjectId
+import pymongo.database
+
+from swh.model.model import Sha1Git
+
+from ..interface import (
+    EntityType,
+    ProvenanceResult,
+    RelationData,
+    RelationType,
+    RevisionData,
+)
+
+
+class ProvenanceStorageMongoDb:
+    def __init__(self, db: pymongo.database.Database):
+        self.db = db
+
+    def content_find_first(self, id: Sha1Git) -> Optional[ProvenanceResult]:
+        pass
+
+    def content_find_all(self, id: Sha1Git, limit: Optional[int] = None) -> Generator[ProvenanceResult, None, None]:
+        pass
+
+    def content_add(self, cnts: Union[Iterable[Sha1Git], Dict[Sha1Git, datetime]]) -> bool:
+        for each_cnt in cnts:
+            try:
+                Entity.factory('content')(each_cnt).add_if_older()
+            except DBError as e:
+                # logging and skipping this item
+                # FIXME, add logging, raise if needed
+                pass
+        return True
+
+    def content_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, datetime]:
+        try:
+            contents = Entity.factory('content').get_all_in_list({'sha1': ids})
+        except DBError as e:
+            # logging and returning None
+            # FIXME, add logging, raise if needed
+            return None
+        return Entity.factory('content').dump_list_as_date_dict(contents)
+
+    def directory_add(self, dirs: Union[Iterable[Sha1Git], Dict[Sha1Git, datetime]]) -> bool:
+        for each_cnt in cnts:
+            try:
+                Entity.factory('directory')(each_cnt).add_if_older()
+            except DBError as e:
+                # logging and skipping this item
+                # FIXME, add logging, raise if needed
+                pass
+        return True
+
+    def directory_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, datetime]:
+        try:
+            directories = Entity.factory('directory').get_all_in_list({'sha1': ids})
+        except DBError as e:
+            # logging and returning None
+            # FIXME, add logging, raise if needed
+            return None
+        return Entity.factory('directory').dump_list_as_date_dict(contents)
+
+    def entity_get_all(self, entity: EntityType) -> Set[Sha1Git]:
+         return Entity.factory(entity).get_all_in_list({'sha1': ids})
+
+    def location_add(self, paths: Iterable[bytes]) -> bool:
+        # TODO: implement this methods if path are to be stored in a separate collection
+        return True
+
+    def location_get_all(self) -> Set[bytes]:
+        pass
+
+    def origin_add(self, orgs: Dict[Sha1Git, str]) -> bool:
+        return True
+
+    def origin_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, str]:
+        pass
+
+    def revision_add(self, revs: Union[Iterable[Sha1Git], Dict[Sha1Git, RevisionData]]) -> bool:
+        return True
+
+    def revision_get(self, ids: Iterable[Sha1Git]) -> Dict[Sha1Git, RevisionData]:
+        pass
+
+    def relation_add(self, relation: RelationType, data: Dict[Sha1Git, Set[RelationData]]) -> bool:
+        return True
+
+    def relation_get(self, relation: RelationType, ids: Iterable[Sha1Git], reverse: bool = False) -> Dict[Sha1Git, Set[RelationData]]:
+        return pass
+
+    def relation_get_all(self, relation: RelationType) -> Dict[Sha1Git, Set[RelationData]]:
+        pass
+
+    def with_path(self) -> bool:
+        return True
diff --git a/swh/provenance/mongo/tests/__init__.py b/swh/provenance/mongo/tests/__init__.py
new file mode 100644
--- /dev/null
+++ b/swh/provenance/mongo/tests/__init__.py
@@ -0,0 +1 @@
+pass