diff --git a/replit_river/codegen/client.py b/replit_river/codegen/client.py
index 20c9cb6..ee1b803 100644
--- a/replit_river/codegen/client.py
+++ b/replit_river/codegen/client.py
@@ -80,6 +80,7 @@
     Literal,
     Optional,
     Mapping,
+    NotRequired,
     Union,
     Tuple,
     TypedDict,
@@ -507,7 +508,9 @@ def extract_props(tpe: RiverType) -> list[dict[str, RiverType]]:
                         encoder_names.add(encoder_name)
                         typeddict_encoder.append(f"{encoder_name}(x[{repr(name)}])")
                         if name not in type.required:
-                            typeddict_encoder.append(f"if x[{repr(name)}] else None")
+                            typeddict_encoder.append(
+                                f"if {repr(name)} in x and x[{repr(name)}] else None"
+                            )
                     elif isinstance(prop, RiverIntersectionType):
                         encoder_name = TypeName(
                             f"encode_{ensure_literal_type(type_name)}"
@@ -541,7 +544,12 @@ def extract_props(tpe: RiverType) -> list[dict[str, RiverType]]:
                             items = cast(RiverConcreteType, prop).items
                             assert items, "Somehow items was none"
                             if is_literal(cast(RiverType, items)):
-                                typeddict_encoder.append(f"x[{repr(name)}]")
+                                if name in prop.required:
+                                    typeddict_encoder.append(f"x[{repr(name)}]")
+                                else:
+                                    typeddict_encoder.append(
+                                        f"x.get({repr(safe_name)})"
+                                    )
                             else:
                                 match type_name:
                                     case ListTypeExpr(inner_type_name):
@@ -606,12 +614,24 @@ def extract_props(tpe: RiverType) -> list[dict[str, RiverType]]:
                         )
                 else:
                     if name not in type.required:
-                        value = ""
-                        if base_model != "TypedDict":
-                            value = " = None"
-                        current_chunks.append(
-                            f"  {name}: Optional[{render_type_expr(type_name)}]{value}"
-                        )
+                        if base_model == "TypedDict":
+                            current_chunks.append(
+                                reindent(
+                                    "  ",
+                                    f"""\
+                        {name}: NotRequired[Optional[{render_type_expr(type_name)}]]
+                                """,
+                                )
+                            )
+                        else:
+                            current_chunks.append(
+                                reindent(
+                                    "  ",
+                                    f"""\
+                        {name}: Optional[{render_type_expr(type_name)}] = None
+                                """,
+                                )
+                            )
                     else:
                         current_chunks.append(
                             f"  {name}: {render_type_expr(type_name)}"