diff --git a/litellm/harness/endpoint.py b/litellm/harness/endpoint.py
index ce3586735c6..21e789ed1fc 100644
--- a/litellm/harness/endpoint.py
+++ b/litellm/harness/endpoint.py
@@ -18,7 +18,7 @@ import secrets
 from collections.abc import AsyncIterable, AsyncIterator, Mapping
 from dataclasses import dataclass
 from types import MappingProxyType, ModuleType
-from typing import TYPE_CHECKING, Any, Final
+from typing import TYPE_CHECKING, Any, Final, Protocol
 
 import httpx
 import openai
@@ -41,7 +41,8 @@ from litellm.types.llms.custom_http import httpxSpecialProvider
 if TYPE_CHECKING:
     from starlette.applications import Starlette
     from starlette.requests import Request
-    from starlette.responses import Response
+    from starlette.responses import JSONResponse, Response, StreamingResponse
+    from starlette.routing import Route
     from uvicorn import Server
 
 verbose_logger: Final = logging.getLogger("LiteLLM")
@@ -85,12 +86,31 @@ COST_HEADER = "x-litellm-response-cost"
 SSE_MEDIA_TYPE = "text/event-stream"
 
 
+class _ApplicationsModule(Protocol):
+    """The starlette.applications attributes the endpoint uses."""
+
+    Starlette: type[Starlette]
+
+
+class _RoutingModule(Protocol):
+    """The starlette.routing attributes the endpoint uses."""
+
+    Route: type[Route]
+
+
+class _ResponsesModule(Protocol):
+    """The starlette.responses attributes the endpoint uses."""
+
+    JSONResponse: type[JSONResponse]
+    StreamingResponse: type[StreamingResponse]
+
+
 @dataclass(frozen=True)
 class _ServerDeps:
     uvicorn: ModuleType
-    applications: ModuleType
-    routing: ModuleType
-    responses: ModuleType
+    applications: _ApplicationsModule
+    routing: _RoutingModule
+    responses: _ResponsesModule
 
 
 def _load_server_deps() -> _ServerDeps:
@@ -193,7 +213,7 @@ class SSEUsageParser:
         if isinstance(event, Mapping):
             self.absorb(event)
 
-    def absorb(self, event: Mapping[str, Any]) -> None:
+    def absorb(self, event: Mapping[str, object]) -> None:
         event_type = event.get("type")
         if event_type == "message_start":
             self._absorb_message_start(event)
@@ -204,16 +224,16 @@ class SSEUsageParser:
         elif isinstance(event.get("usage"), Mapping):
             self._set(*usage_from_mapping(event["usage"]))
 
-    def _absorb_message_start(self, event: Mapping[str, Any]) -> None:
+    def _absorb_message_start(self, event: Mapping[str, object]) -> None:
         message = event.get("message")
         if isinstance(message, Mapping):
             self._set(*usage_from_mapping(message.get("usage")))
 
-    def _absorb_message_delta(self, event: Mapping[str, Any]) -> None:
+    def _absorb_message_delta(self, event: Mapping[str, object]) -> None:
         # message_delta output_tokens is cumulative for the whole message.
         self._set(*usage_from_mapping(event.get("usage")))
 
-    def _absorb_response_completed(self, event: Mapping[str, Any]) -> None:
+    def _absorb_response_completed(self, event: Mapping[str, object]) -> None:
         response = event.get("response")
         if isinstance(response, Mapping):
             self._set(*usage_from_mapping(response.get("usage")))
@@ -271,7 +291,7 @@ def gateway_headers(
     incoming: Mapping[str, str],
     gateway: GatewayTarget,
     harness: Harness,
-    metadata: Mapping[str, Any] | None,
+    metadata: Mapping[str, object] | None,
 ) -> Mapping[str, str]:
     """Incoming headers minus hop-by-hop/auth/x-litellm-*, plus gateway auth, tags, metadata."""
     kept = (
@@ -309,7 +329,7 @@ def error_status(exc: BaseException) -> int:
     return 500
 
 
-def error_body(exc: BaseException, message: str) -> dict[str, Any]:  # mutable-ok: JSONResponse body
+def error_body(exc: BaseException, message: str) -> dict[str, dict[str, str]]:  # mutable-ok: JSONResponse body
     return {"error": {"type": type(exc).__name__, "message": message}}  # mutable-ok: JSONResponse body
 
 
@@ -373,7 +393,7 @@ class ModelEndpoint:
         gateway: GatewayTarget | None,
         api_key: str | None = None,
         api_base: str | None = None,
-        metadata: Mapping[str, Any] | None = None,
+        metadata: Mapping[str, object] | None = None,
         *,
         client: httpx.AsyncClient | None = None,
     ) -> None:
@@ -382,14 +402,14 @@ class ModelEndpoint:
         self.gateway = gateway
         self.api_key = api_key
         self.api_base = api_base
-        self.metadata: Mapping[str, Any] = MappingProxyType(dict(metadata or ()))
+        self.metadata: Mapping[str, object] = MappingProxyType(dict(metadata or ()))
         self.token = secrets.token_urlsafe(HARNESS_SESSION_TOKEN_BYTES)
         self.usage = UsageTracker()
         self.port = 0
         self._injected_client = client
         self._deps: _ServerDeps | None = None
         self._client: httpx.AsyncClient | None = None
-        self._server: Any = None
+        self._server: Server | None = None
         self._task: asyncio.Task[None] | None = None
 
     @property
@@ -410,7 +430,7 @@ class ModelEndpoint:
         self._server = self._build_server(self._deps)
         self._task = asyncio.create_task(self._server.serve())
         try:
-            await asyncio.wait_for(self._wait_started(), HARNESS_ENDPOINT_STARTUP_TIMEOUT_SECONDS)
+            await asyncio.wait_for(self._wait_started(self._server), HARNESS_ENDPOINT_STARTUP_TIMEOUT_SECONDS)
         except BaseException:
             await self.stop()
             raise
@@ -439,8 +459,8 @@ class ModelEndpoint:
         )
         return handler.client
 
-    async def _wait_started(self) -> None:
-        while not self._server.started:
+    async def _wait_started(self, server: Server) -> None:
+        while not server.started:
             if self._task is not None and self._task.done():
                 raise HarnessError("harness model endpoint failed to start")
             await asyncio.sleep(DEFAULT_POLLING_INTERVAL)
@@ -483,7 +503,7 @@ class ModelEndpoint:
         )
 
     @property
-    def _responses(self) -> ModuleType:
+    def _responses(self) -> _ResponsesModule:
         if self._deps is None:
             raise HarnessError("harness model endpoint is not started")
         return self._deps.responses
@@ -530,7 +550,7 @@ class ModelEndpoint:
             return await self._forward(request, route, body)
         return await self._call_sdk(route, body)
 
-    def _cost_model(self, body: Mapping[str, Any]) -> str | None:
+    def _cost_model(self, body: Mapping[str, object]) -> str | None:
         model = self.model or body.get("model")
         return model if isinstance(model, str) else None
 
@@ -545,7 +565,7 @@ class ModelEndpoint:
             cost = compute_cost(model, input_tokens, output_tokens)
         self.usage.add(input_tokens, output_tokens, cost)
 
-    async def _forward(self, request: Request, route: str, body: Mapping[str, Any]) -> Response:
+    async def _forward(self, request: Request, route: str, body: Mapping[str, object]) -> Response:
         if self._client is None or self.gateway is None:
             raise HarnessError("gateway client is not started")
         if self.model:
@@ -601,7 +621,7 @@ class ModelEndpoint:
         self._record(model, tokens[0], tokens[1], header_cost(upstream.headers))
 
     def _sdk_kwargs(
-        self, body: Mapping[str, Any]
+        self, body: Mapping[str, object]
     ) -> dict[str, Any]:  # mutable-ok: SDK call kwargs, mutated by _invoke_sdk then splatted
         kwargs: dict[str, Any] = {**body}  # mutable-ok: SDK call kwargs built from the JSON body, then overridden
         if self.model:
@@ -629,7 +649,7 @@ class ModelEndpoint:
             return await litellm.acompletion(**kwargs)
         return await litellm.aresponses(**kwargs)
 
-    async def _call_sdk(self, route: str, body: Mapping[str, Any]) -> Response:
+    async def _call_sdk(self, route: str, body: Mapping[str, object]) -> Response:
         kwargs = self._sdk_kwargs(body)
         model = self._cost_model(kwargs)
         try:
diff --git a/litellm/harness/runtime.py b/litellm/harness/runtime.py
index 4ec167a06b4..da90b28a751 100644
--- a/litellm/harness/runtime.py
+++ b/litellm/harness/runtime.py
@@ -79,7 +79,7 @@ class SessionConfig:
     api_key: str | None = None
     api_base: str | None = None
     instructions: str | None = None
-    tools: Sequence[Callable[..., Any]] = ()
+    tools: Sequence[Callable[..., object]] = ()
     skills: Sequence[str] = ()
     disable_tools: Sequence[str] = ()
     permissions: PermissionMode = "full"
@@ -87,7 +87,7 @@ class SessionConfig:
     output: type[BaseModel] | None = None
     max_turns: int | None = None
     timeout: float | None = None
-    metadata: Mapping[str, Any] = field(default_factory=dict)
+    metadata: Mapping[str, object] = field(default_factory=dict)
     options: HarnessOptions | None = None
     install: bool = False
 
@@ -184,7 +184,7 @@ def build_config(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -192,7 +192,7 @@ def build_config(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> SessionConfig:
@@ -335,7 +335,7 @@ async def call_approval_handler(handler: ApprovalHandler, approval: Approval) ->
     """Run on_approval (sync in a worker thread, or async) and resolve approval."""
     try:
         if inspect.iscoroutinefunction(handler):
-            decision: Any = await handler(approval)
+            decision: object = await handler(approval)
         else:
             decision = await asyncio.to_thread(handler, approval)
             if inspect.isawaitable(decision):
@@ -794,7 +794,7 @@ def aagent_session(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -802,7 +802,7 @@ def aagent_session(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> AsyncSession:
@@ -838,7 +838,7 @@ async def arun_agent(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -846,7 +846,7 @@ async def arun_agent(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> Result:
@@ -883,7 +883,7 @@ def astream_agent(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -891,7 +891,7 @@ def astream_agent(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> AsyncEventStream:
@@ -935,7 +935,7 @@ def aagent_resume(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -943,7 +943,7 @@ def aagent_resume(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> AsyncSession:
@@ -988,7 +988,7 @@ def aagent(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -996,10 +996,10 @@ def aagent(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
-) -> Coroutine[Any, Any, Result] | AsyncEventStream:
+) -> Coroutine[object, object, Result] | AsyncEventStream:
     """Run an agent harness on one prompt.
 
     `await litellm.aagent(...)` returns a Result. With stream=True it returns an async
diff --git a/litellm/harness/sync.py b/litellm/harness/sync.py
index 1788543f9b5..1cb8fbcbca2 100644
--- a/litellm/harness/sync.py
+++ b/litellm/harness/sync.py
@@ -62,7 +62,7 @@ class _LoopThread:
                 self._thread.start()
             return self._loop
 
-    def submit(self, coro: Coroutine[Any, Any, T]) -> Future[T]:
+    def submit(self, coro: Coroutine[object, None, T]) -> Future[T]:
         return asyncio.run_coroutine_threadsafe(coro, self.loop())
 
 
@@ -77,7 +77,7 @@ def _ensure_sync_context(name: str) -> None:
     raise RuntimeError(IN_LOOP_MESSAGE.format(name=name))
 
 
-def run_sync(coro: Coroutine[Any, Any, T], name: str) -> T:
+def run_sync(coro: Coroutine[object, None, T], name: str) -> T:
     """Run coro on the harness loop thread and block for its result."""
     try:
         _ensure_sync_context(name)
@@ -215,7 +215,7 @@ def _run(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -223,7 +223,7 @@ def _run(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> Result:
@@ -262,7 +262,7 @@ def _stream(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -270,7 +270,7 @@ def _stream(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> EventStream:
@@ -307,7 +307,7 @@ def agent_session(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -315,7 +315,7 @@ def agent_session(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> Session:
@@ -352,7 +352,7 @@ def agent_resume(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -360,7 +360,7 @@ def agent_resume(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> Session:
@@ -399,7 +399,7 @@ def agent(
     api_key: str | None = None,
     api_base: str | None = None,
     instructions: str | None = None,
-    tools: Sequence[Callable[..., Any]] = (),
+    tools: Sequence[Callable[..., object]] = (),
     skills: Sequence[str | os.PathLike[str]] = (),
     disable_tools: Sequence[str] = (),
     permissions: PermissionMode = "full",
@@ -407,7 +407,7 @@ def agent(
     output: type[BaseModel] | None = None,
     max_turns: int | None = None,
     timeout: float | None = None,
-    metadata: Mapping[str, Any] | None = None,
+    metadata: Mapping[str, object] | None = None,
     options: HarnessOptions | None = None,
     install: bool = False,
 ) -> Result | EventStream:
diff --git a/litellm/responses/main.py b/litellm/responses/main.py
index d145f8cc6b8..14c12fc571d 100644
--- a/litellm/responses/main.py
+++ b/litellm/responses/main.py
@@ -994,7 +994,12 @@ def _responses_try_dispatch_mcp_gateway(
     kwargs: dict[str, object],
     _is_async: bool,
     skip_mcp_handler: bool,
-) -> Any | None:
+) -> (
+    ResponsesAPIResponse
+    | BaseResponsesAPIStreamingIterator
+    | Coroutine[object, object, ResponsesAPIResponse | BaseResponsesAPIStreamingIterator]
+    | None
+):
     """Return a response when MCP gateway handles the call; otherwise None."""
     from litellm.responses.mcp.litellm_proxy_mcp_handler import (
         LiteLLM_Proxy_MCP_Handler,