diff --git a/python/mujoco_mpc/mjx/README.md b/python/mujoco_mpc/mjx/README.md
new file mode 100644
index 000000000..b4c82a32b
--- /dev/null
+++ b/python/mujoco_mpc/mjx/README.md
@@ -0,0 +1,10 @@
+# MJX Predictive Sampling
+
+Run `handover` example:
+
+```sh
+python visualize.py
+```
+
+##
+Requires: mujoco, mujoco-mjx, jax[cuda], matplotlib, mediapy (Python), ffmpeg
\ No newline at end of file
diff --git a/python/mujoco_mpc/mjx/tasks/bimanual/handover.py b/python/mujoco_mpc/mjx/tasks/bimanual/handover.py
index 4d1d81205..31efb3f97 100644
--- a/python/mujoco_mpc/mjx/tasks/bimanual/handover.py
+++ b/python/mujoco_mpc/mjx/tasks/bimanual/handover.py
@@ -13,13 +13,14 @@
 # limitations under the License.
 # ==============================================================================
 
-from etils import epath
+from typing import Callable
+from pathlib import Path
 import jax
 from jax import numpy as jp
 import mujoco
 from mujoco import mjx
-from mujoco_mpc.mjx import predictive_sampling
 
+CostFn = Callable[[mjx.Model, mjx.Data], jax.Array]
 
 def bring_to_target(m: mjx.Model, d: mjx.Data) -> jax.Array:
   """Returns cost for bimanual bring to target task."""
@@ -48,22 +49,14 @@ def bring_to_target(m: mjx.Model, d: mjx.Data) -> jax.Array:
 
 
 def get_models_and_cost_fn() -> (
-    tuple[mujoco.MjModel, mujoco.MjModel, predictive_sampling.CostFn]
+    tuple[mujoco.MjModel, mujoco.MjModel, CostFn]
 ):
   """Returns a tuple of the model and the cost function."""
-  path = epath.Path(
-      'build/mjpc/tasks/bimanual/'
+  model_path = (
+      Path(__file__).parent.parent.parent
+      / "../../../build/mjpc/tasks/bimanual/mjx_scene.xml"
   )
-  model_file_name = 'mjx_scene.xml'
-  xml = (path / model_file_name).read_text()
-  assets = {}
-  for f in path.glob('*.xml'):
-    if f.name == model_file_name:
-      continue
-    assets[f.name] = f.read_bytes()
-  for f in (path / 'assets').glob('*'):
-    assets[f.name] = f.read_bytes()
-  sim_model = mujoco.MjModel.from_xml_string(xml, assets)
-  plan_model = mujoco.MjModel.from_xml_string(xml, assets)
+  sim_model = mujoco.MjModel.from_xml_path(str(model_path))
+  plan_model = mujoco.MjModel.from_xml_path(str(model_path))
   plan_model.opt.timestep = 0.01  # incidentally, already the case
   return sim_model, plan_model, bring_to_target
diff --git a/python/mujoco_mpc/mjx/visualize.py b/python/mujoco_mpc/mjx/visualize.py
index deb493d65..b3b21af87 100644
--- a/python/mujoco_mpc/mjx/visualize.py
+++ b/python/mujoco_mpc/mjx/visualize.py
@@ -16,8 +16,8 @@
 import matplotlib.pyplot as plt
 import mediapy
 import mujoco
-from mujoco_mpc.mjx import predictive_sampling
-from mujoco_mpc.mjx.tasks.bimanual import handover
+import predictive_sampling
+from tasks.bimanual import handover
 # %%
 nsteps = 500
 steps_per_plan = 4