diff --git a/sdk/python/README.md b/sdk/python/README.md
index 4f150bb692..0c5a8bd3a6 100644
--- a/sdk/python/README.md
+++ b/sdk/python/README.md
@@ -3,9 +3,8 @@
 Experimental Python SDK for `codex app-server` JSON-RPC v2 over stdio, with a small default surface optimized for real scripts and apps.
 
 The generated wire-model layer is sourced from the pinned `openai-codex-cli-bin`
-runtime package. The package root exports the supported ergonomic SDK surface
-plus selected public value types; generated transport modules are shipped for
-runtime use but are not the supported public API.
+runtime package and exposed as Pydantic models with snake_case Python fields
+that serialize back to the app-server’s camelCase wire format.
 
 ## Install
 
@@ -111,4 +110,4 @@ This supports the CI release flow:
 - Use context managers (`with Codex() as codex:`) to ensure shutdown.
 - Prefer `thread.run("...")` for the common case. Use `thread.turn(...)` when
   you need streaming, steering, or interrupt control.
-- For transient overload, import `retry_on_overload` from `codex_app_server`.
+- For transient overload, use `codex_app_server.retry.retry_on_overload`.
diff --git a/sdk/python/docs/api-reference.md b/sdk/python/docs/api-reference.md
index 8793929dcd..c7a763498d 100644
--- a/sdk/python/docs/api-reference.md
+++ b/sdk/python/docs/api-reference.md
@@ -10,12 +10,12 @@ This SDK surface is experimental. Turn streams are routed by turn ID so one clie
 from codex_app_server import (
     Codex,
     AsyncCodex,
-    AppServerConfig,
     RunResult,
     Thread,
     AsyncThread,
     TurnHandle,
     AsyncTurnHandle,
+    InitializeResponse,
     Input,
     InputItem,
     TextInput,
@@ -24,14 +24,13 @@ from codex_app_server import (
     SkillInput,
     MentionInput,
     TurnStatus,
-    retry_on_overload,
 )
+from codex_app_server.generated.v2_all import ThreadItem, ThreadTokenUsage
 ```
 
 - Version: `codex_app_server.__version__`
 - Requires Python >= 3.10
-- The package root `__all__` is the supported public import surface; generated
-  transport modules are shipped for runtime use but are not re-exported.
+- Canonical generated app-server models live in `codex_app_server.generated.v2_all`
 
 ## Codex (sync)
 
@@ -166,23 +165,20 @@ InputItem = TextInput | ImageInput | LocalImageInput | SkillInput | MentionInput
 Input = list[InputItem] | InputItem
 ```
 
-## Public Value Types
+## Generated Models
 
-The package root exports the ergonomic SDK plus selected value types used by
-public method signatures:
+The SDK wrappers return and accept canonical generated app-server models wherever possible:
 
 ```python
-from codex_app_server import (
+from codex_app_server.generated.v2_all import (
     AskForApproval,
-    ReasoningEffort,
-    SandboxPolicy,
+    ThreadReadResponse,
+    Turn,
+    TurnStartParams,
     TurnStatus,
 )
 ```
 
-Generated transport modules are shipped for SDK runtime use, but they are not
-the supported root API.
-
 ## Retry + errors
 
 ```python
diff --git a/sdk/python/docs/getting-started.md b/sdk/python/docs/getting-started.md
index 6d5e4ab26c..1794d39f70 100644
--- a/sdk/python/docs/getting-started.md
+++ b/sdk/python/docs/getting-started.md
@@ -95,12 +95,12 @@ with Codex() as codex:
     print(result.final_response)
 ```
 
-## 6) Public value types
+## 6) Generated models
 
-The convenience wrappers and selected value types live at the package root:
+The convenience wrappers live at the package root, but the canonical app-server models live under:
 
 ```python
-from codex_app_server import AskForApproval, ReasoningEffort, SandboxPolicy, TurnStatus
+from codex_app_server.generated.v2_all import Turn, TurnStatus, ThreadReadResponse
 ```
 
 ## 7) Next stops
diff --git a/sdk/python/examples/11_cli_mini_app/async.py b/sdk/python/examples/11_cli_mini_app/async.py
index 2e0d2acba0..4216cf7820 100644
--- a/sdk/python/examples/11_cli_mini_app/async.py
+++ b/sdk/python/examples/11_cli_mini_app/async.py
@@ -14,6 +14,8 @@ import asyncio
 from codex_app_server import (
     AsyncCodex,
     TextInput,
+    ThreadTokenUsageUpdatedNotification,
+    TurnCompletedNotification,
 )
 
 
@@ -70,13 +72,12 @@ async def main() -> None:
                         print(delta, end="", flush=True)
                         printed_delta = True
                     continue
-                if event.method == "thread/tokenUsage/updated":
-                    usage = getattr(payload, "token_usage", None)
+                if isinstance(payload, ThreadTokenUsageUpdatedNotification):
+                    usage = payload.token_usage
                     continue
-                if event.method == "turn/completed":
-                    turn = getattr(payload, "turn", None)
-                    status = getattr(turn, "status", None)
-                    error = getattr(turn, "error", None)
+                if isinstance(payload, TurnCompletedNotification):
+                    status = payload.turn.status
+                    error = payload.turn.error
 
             if printed_delta:
                 print()
diff --git a/sdk/python/examples/11_cli_mini_app/sync.py b/sdk/python/examples/11_cli_mini_app/sync.py
index 7b4c37cb23..e961cfbcc3 100644
--- a/sdk/python/examples/11_cli_mini_app/sync.py
+++ b/sdk/python/examples/11_cli_mini_app/sync.py
@@ -12,6 +12,8 @@ ensure_local_sdk_src()
 from codex_app_server import (
     Codex,
     TextInput,
+    ThreadTokenUsageUpdatedNotification,
+    TurnCompletedNotification,
 )
 
 print("Codex mini CLI. Type /exit to quit.")
@@ -67,13 +69,12 @@ with Codex(config=runtime_config()) as codex:
                     print(delta, end="", flush=True)
                     printed_delta = True
                 continue
-            if event.method == "thread/tokenUsage/updated":
-                usage = getattr(payload, "token_usage", None)
+            if isinstance(payload, ThreadTokenUsageUpdatedNotification):
+                usage = payload.token_usage
                 continue
-            if event.method == "turn/completed":
-                turn = getattr(payload, "turn", None)
-                status = getattr(turn, "status", None)
-                error = getattr(turn, "error", None)
+            if isinstance(payload, TurnCompletedNotification):
+                status = payload.turn.status
+                error = payload.turn.error
 
         if printed_delta:
             print()
diff --git a/sdk/python/src/codex_app_server/__init__.py b/sdk/python/src/codex_app_server/__init__.py
index c950588325..49753f4e24 100644
--- a/sdk/python/src/codex_app_server/__init__.py
+++ b/sdk/python/src/codex_app_server/__init__.py
@@ -24,12 +24,22 @@ from .generated.v2_all import (
     SandboxPolicy,
     SortDirection,
     ThreadListCwdFilter,
+    ThreadItem,
+    ThreadForkParams,
+    ThreadListParams,
+    ThreadResumeParams,
     ThreadSortKey,
     ThreadSource,
     ThreadSourceKind,
+    ThreadStartParams,
     ThreadStartSource,
+    ThreadTokenUsageUpdatedNotification,
+    TurnCompletedNotification,
+    TurnStartParams,
     TurnStatus,
+    TurnSteerParams,
 )
+from .models import InitializeResponse
 from .api import (
     AsyncCodex,
     AsyncThread,
@@ -58,6 +68,7 @@ __all__ = [
     "AsyncThread",
     "TurnHandle",
     "AsyncTurnHandle",
+    "InitializeResponse",
     "RunResult",
     "Input",
     "InputItem",
@@ -66,6 +77,9 @@ __all__ = [
     "LocalImageInput",
     "SkillInput",
     "MentionInput",
+    "ThreadItem",
+    "ThreadTokenUsageUpdatedNotification",
+    "TurnCompletedNotification",
     "ApprovalsReviewer",
     "AskForApproval",
     "Personality",
@@ -76,11 +90,17 @@ __all__ = [
     "SandboxPolicy",
     "SortDirection",
     "ThreadListCwdFilter",
+    "ThreadStartParams",
+    "ThreadResumeParams",
+    "ThreadListParams",
     "ThreadSortKey",
     "ThreadSource",
     "ThreadSourceKind",
+    "ThreadForkParams",
     "ThreadStartSource",
     "TurnStatus",
+    "TurnStartParams",
+    "TurnSteerParams",
     "retry_on_overload",
     "AppServerError",
     "TransportClosedError",
diff --git a/sdk/python/tests/test_public_api_signatures.py b/sdk/python/tests/test_public_api_signatures.py
index 7ae5c706e9..2a956256d2 100644
--- a/sdk/python/tests/test_public_api_signatures.py
+++ b/sdk/python/tests/test_public_api_signatures.py
@@ -26,6 +26,7 @@ EXPECTED_ROOT_EXPORTS = [
     "AsyncThread",
     "TurnHandle",
     "AsyncTurnHandle",
+    "InitializeResponse",
     "RunResult",
     "Input",
     "InputItem",
@@ -34,6 +35,9 @@ EXPECTED_ROOT_EXPORTS = [
     "LocalImageInput",
     "SkillInput",
     "MentionInput",
+    "ThreadItem",
+    "ThreadTokenUsageUpdatedNotification",
+    "TurnCompletedNotification",
     "ApprovalsReviewer",
     "AskForApproval",
     "Personality",
@@ -44,11 +48,17 @@ EXPECTED_ROOT_EXPORTS = [
     "SandboxPolicy",
     "SortDirection",
     "ThreadListCwdFilter",
+    "ThreadStartParams",
+    "ThreadResumeParams",
+    "ThreadListParams",
     "ThreadSortKey",
     "ThreadSource",
     "ThreadSourceKind",
+    "ThreadForkParams",
     "ThreadStartSource",
     "TurnStatus",
+    "TurnStartParams",
+    "TurnSteerParams",
     "retry_on_overload",
     "AppServerError",
     "TransportClosedError",
@@ -115,15 +125,9 @@ def test_package_root_exports_only_public_api() -> None:
     assert {
         "AppServerClient": hasattr(codex_app_server, "AppServerClient"),
         "AsyncAppServerClient": hasattr(codex_app_server, "AsyncAppServerClient"),
-        "InitializeResponse": hasattr(codex_app_server, "InitializeResponse"),
-        "ThreadStartParams": hasattr(codex_app_server, "ThreadStartParams"),
-        "TurnStartParams": hasattr(codex_app_server, "TurnStartParams"),
     } == {
         "AppServerClient": False,
         "AsyncAppServerClient": False,
-        "InitializeResponse": False,
-        "ThreadStartParams": False,
-        "TurnStartParams": False,
     }