diff --git a/src/cript/nodes/main.py b/src/cript/nodes/main.py
index db95ecc..0eecdca 100644
--- a/src/cript/nodes/main.py
+++ b/src/cript/nodes/main.py
@@ -7,9 +7,11 @@
 from jsonschema.exceptions import best_match
 from uuid import uuid4
 
+import cript
 from cript import Cript, NotFoundError, camel_case_to_snake_case, extract_node_from_result
 from .schema import cript_schema
 
+
 logger = logging.getLogger(__name__)
 
 
@@ -218,7 +220,13 @@ def __getattr__(self, key):
         try:
             return self.__getitem__(key)
         except KeyError:
-            raise AttributeError(key)
+            # TODO consider a caching of these paginators
+            if key in self.children:
+                child_paginator = cript.resources.child.ChildPaginator(self, key)
+                self.__dict__[key] = child_paginator
+                return self.__dict__[key]
+            else:
+                raise AttributeError(key)
 
     def __setattr__(self, key, value):
         self.__setitem__(key, value)
@@ -351,7 +359,7 @@ def retrieve_child(self, parent, child):
         try:
             if not self._primary_key:
                 result = self.__dict__["client"].nodes.retrieve_children(
-                    node=parent.name_url, uuid=parent.uuid, child_node=child.name_url
+                    node=parent.name_url, uuid=parent.uuid, child_node=child.name_url,
                 )
             else:
                 result = self.__dict__["client"].search.exact.child_node(
diff --git a/src/cript/resources/__init__.py b/src/cript/resources/__init__.py
index 45d9905..61e3032 100644
--- a/src/cript/resources/__init__.py
+++ b/src/cript/resources/__init__.py
@@ -33,6 +33,10 @@
     AsyncControlledVocabulariesResourceWithStreamingResponse,
 )
 
+from .child import (
+    ChildPaginator,
+)
+
 __all__ = [
     "SchemaResource",
     "AsyncSchemaResource",
@@ -58,4 +62,5 @@
     "AsyncControlledVocabulariesResourceWithRawResponse",
     "ControlledVocabulariesResourceWithStreamingResponse",
     "AsyncControlledVocabulariesResourceWithStreamingResponse",
+    "ChildPaginator",
 ]
diff --git a/src/cript/resources/child.py b/src/cript/resources/child.py
new file mode 100644
index 0000000..147aec1
--- /dev/null
+++ b/src/cript/resources/child.py
@@ -0,0 +1,100 @@
+import httpx
+from .._types import NOT_GIVEN, Body, Query, Headers, NotGiven
+from .._utils import (
+    maybe_transform,
+    async_maybe_transform,
+)
+from .._compat import cached_property
+from .._response import (
+    to_raw_response_wrapper,
+    to_streamed_response_wrapper,
+    async_to_raw_response_wrapper,
+    async_to_streamed_response_wrapper,
+)
+from .._base_client import (
+    make_request_options,
+)
+from ..types.shared.search import Search
+from .._resource import SyncAPIResource
+
+class ChildPaginator:
+    # TODO consider writing operations
+    def __init__(self, parent, child, client=None):
+        if client is None:
+            client = parent.client
+        self._client = client
+        self._parent = parent
+        self._child = child
+
+        self._current_child_list = []
+        self._current_child_position = 0
+        # TODO change to after
+        self._current_page = 0
+        self._count = None
+
+    def __iter__(self):
+        self._current_child_position = 0
+        return self
+
+    def __next__(self):
+        if self._current_child_position >= len(self._current_child_list):
+            self._fetch_next_page()
+        try:
+            next_node = self._current_child_list[self._current_child_position]
+        except IndexError:
+            raise StopIteration
+
+        self._current_child_position += 1
+
+        return next_node
+
+    def _fetch_next_page(self):
+        if self._finished_fetching:
+            raise StopIteration
+
+        # TODO change to after
+        response = self._client.nodes.retrieve_children(uuid=self._parent.uuid, node=self._parent.name_url, child_node=self._child, page=self._current_page)
+        self._current_page += 1
+
+        if self._count is not None and self._count != int(response.data.count):
+            raise RuntimeError("The number of elements for a child iteration changed during pagination. This may lead to inconsistencies. Please try again.")
+        self._count = int(response.data.count)
+
+        self._current_child_list += response.data.result
+
+    # Make it a random access iterator, since ppl expect it to behave list a list
+    def __getitem__(self, key):
+        key_index = int(key)
+        previous_pos = self._current_child_position
+        try:
+            if key_index < 0:
+                while not self._finished_fetching:
+                    next(self)
+
+            while len(self._current_child_list) <= key_index:
+                try:
+                    next(self)
+                except StopIteration:
+                    break
+        finally:
+            self._current_child_position = previous_pos
+        # We don't need explicit bounds checking, since the list access does that for us.
+        return self._current_child_list[key_index]
+
+    def __len__(self):
+        previous_pos = self._current_child_position
+        try:
+            if self._count is None:
+                try:
+                    next(iter(self))
+                except StopIteration:
+                    self._count = 0
+        finally:
+            self._current_child_position = previous_pos
+        return self._count
+
+    @property
+    def _finished_fetching(self):
+        if self._count is None:
+            return False
+        return len(self._current_child_list) == self._count
diff --git a/src/cript/resources/nodes.py b/src/cript/resources/nodes.py
index 3ef9e46..0cbbccf 100644
--- a/src/cript/resources/nodes.py
+++ b/src/cript/resources/nodes.py
@@ -151,6 +151,8 @@ def retrieve_children(
         *,
         node: str,
         child_node: str,
+        # TODO change to after
+        page: int | None = None,
         # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
         # The extra values given here take precedence over values defined on the client or passed to this method.
         extra_headers: Headers | None = None,
@@ -176,10 +178,19 @@ def retrieve_children(
             raise ValueError(f"Expected a non-empty value for `uuid` but received {uuid!r}")
         if not child_node:
             raise ValueError(f"Expected a non-empty value for `child_node` but received {child_node!r}")
+        # TODO change to after
+        if page is not None:
+            query = {"page": page}
+        else:
+            query = {} # Does it make sense to allow non-paginated retrieval? The current Code uses it.
         return self._get(
             f"/{node}/{uuid}/{child_node}",
             options=make_request_options(
-                extra_headers=extra_headers, extra_query=extra_query, extra_body=extra_body, timeout=timeout
+                extra_headers=extra_headers,
+                extra_query=extra_query,
+                extra_body=extra_body,
+                timeout=timeout,
+                query=query,
             ),
             cast_to=Search,
         )
diff --git a/tests/api_resources/test_cript.py b/tests/api_resources/test_cript.py
index b1c2cde..8ef50f2 100644
--- a/tests/api_resources/test_cript.py
+++ b/tests/api_resources/test_cript.py
@@ -7,6 +7,7 @@
 
 import pytest
 
+import cript
 from cript import *
 
 base_url = os.environ.get("TEST_API_BASE_URL", "http://127.0.0.1:4010")
@@ -48,7 +49,7 @@ def test_create_project(self) -> None:
             notes="my notes",
         )
         assert node.get("name") is not None
-    
+
     def test_create_collection_exisiting_project(self) -> None:
         col1=Collection(name=generic_collection)
         proj = Project(uuid=CREATED_UUID, collection=[col1])
@@ -66,6 +67,9 @@ def test_create_experiment_exisiting_collection(self) -> None:
         col1 = Collection(name=generic_collection, experiment=[exp1])
         proj1 = Project(uuid=CREATED_UUID, collection=[col1])
         assert exp1.get("name") == generic_experiment
+        assert proj1.collection[0].get("name") == col1.name
+        # TODO full node access
+        # assert proj1.collection[0].experiment[0].name == exp1.name
 
 
     def test_create_material(self) -> None:
@@ -246,6 +250,63 @@ def test_unlink_all_children(self) -> None:
         proj1.delete(material=None)
         assert proj1.get("material") is None
 
+    def test_child_paginator(self)->None:
+        material_list = []
+        num_mat = 15
+        for i in range(num_mat):
+            mat1 = Material(
+                name=f"{generic_material1} #{i}",
+                bigsmiles="{[][<]CCO[>][]}",
+            )
+            material_list += [mat1]
+        proj1 = Project(uuid=CREATED_UUID, material=material_list)
+
+        paginator_iter = proj1.material
+        for i, child in enumerate(paginator_iter):
+            assert child.get("name").endswith(f"#{i}")
+
+        paginator_len = cript.resources.child.ChildPaginator(proj1, "material")
+        # First time is doing it on empty
+        assert len(paginator_len) == num_mat
+        # Second time, it should have fetched
+        assert len(paginator_len) == num_mat
+
+        paginator_rand = cript.resources.child.ChildPaginator(proj1, "material")
+
+        # Accessing the second page right away
+        idx = 12
+        child = paginator_rand[idx]
+        assert child.get("name").endswith(f"#{idx}")
+        # And again, shouldn't trigger a new fetch
+        idx = 13
+        child = paginator_rand[idx]
+        assert child.get("name").endswith(f"#{idx}")
+
+        with pytest.raises(IndexError):
+            paginator_rand[num_mat+4]
+
+        paginator_neg = cript.resources.child.ChildPaginator(proj1, "material")
+
+        # Accessing the second page right away
+        idx = -1
+        child = paginator_neg[idx]
+        assert child.get("name").endswith(f"#{num_mat-1}")
+
+        # Test list conversion
+        paginator_list = cript.resources.child.ChildPaginator(proj1, "material")
+        fetched_material_list = list(paginator_list)
+        for i, child in enumerate(fetched_material_list):
+            assert child.get("name").endswith(f"#{i}")
+
+        # Test empty paginator
+        paginator_empty = cript.resources.child.ChildPaginator(proj1, "inventory")
+        assert len(paginator_empty) == 0
+
+        # Test empty paginator
+        paginator_non_exist = cript.resources.child.ChildPaginator(proj1, "non-existent attribute")
+        with pytest.raises(cript.NotFoundError):
+            len(paginator_non_exist)
+
     def test_delete_node(self) -> None:
         proj1 = Project(uuid=CREATED_UUID)
         proj1.delete()