diff --git a/pgx/_src/games/shogi.py b/pgx/_src/games/shogi.py
index d17149e52..3583e84b2 100644
--- a/pgx/_src/games/shogi.py
+++ b/pgx/_src/games/shogi.py
@@ -80,6 +80,20 @@ class GameState(NamedTuple):
     cache_king: Array = jnp.int32(44)
 
 
+class Game:
+    def init(self) -> GameState:
+        return GameState()
+
+    def step(self, state: GameState, action: Array) -> GameState:
+        return _step(state, action)
+
+    def observe(self, state: GameState) -> Array:
+        return _observe(state, False)
+
+    def legal_action_mask(self, state: GameState) -> Array:
+        return _legal_action_mask(state)
+
+
 @dataclass
 class Action:
     is_drop: Array
diff --git a/pgx/shogi.py b/pgx/shogi.py
index 2d99ec1ff..57c29fdad 100644
--- a/pgx/shogi.py
+++ b/pgx/shogi.py
@@ -24,7 +24,7 @@
 )
 from pgx._src.struct import dataclass
 from pgx._src.types import Array, PRNGKey
-from pgx._src.games.shogi import MAX_TERMINATION_STEPS, GameState, _step, _legal_action_mask, _observe, _flip
+from pgx._src.games.shogi import MAX_TERMINATION_STEPS, GameState, Game, _observe, _flip
 
 
 TRUE = jnp.bool_(True)
@@ -54,7 +54,7 @@ def _from_board(turn, piece_board: Array, hand: Array):
         # fmt: off
         state = jax.lax.cond(turn % 2 == 1, lambda: state.replace(_x=_flip(state._x)), lambda: state)  # type: ignore
         # fmt: on
-        return state.replace(legal_action_mask=_legal_action_mask(state._x))  # type: ignore
+        return state.replace(legal_action_mask=Game().legal_action_mask(state._x))  # type: ignore
 
     @staticmethod
     def _from_sfen(sfen):
@@ -67,8 +67,10 @@ def _to_sfen(self):
 
 
 class Shogi(core.Env):
+
     def __init__(self):
         super().__init__()
+        self._game = Game()
 
     def _init(self, key: PRNGKey) -> State:
         state = State()
@@ -79,13 +81,13 @@ def _step(self, state: core.State, action: Array, key) -> State:
         del key
         assert isinstance(state, State)
         # Note: Assume that illegal action is already filtered by Env.step
-        x = _step(state._x, action)
+        x = self._game.step(state._x, action)
         state = state.replace(  # type: ignore
             current_player=(state.current_player + 1) % 2,
             _x=x,
         )
         del x
-        legal_action_mask = _legal_action_mask(state._x)
+        legal_action_mask = self._game.legal_action_mask(state._x)
         terminated = ~legal_action_mask.any()
         # fmt: off
         reward = jax.lax.select(