llama-stack 0.3.5__py3-none-any.whl → 0.4.0__py3-none-any.whl

llama_stack_api/inspect_api/fastapi_routes.py

@@ -0,0 +1,76 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+"""FastAPI router for the Inspect API.
+
+This module defines the FastAPI router for the Inspect API using standard
+FastAPI route decorators.
+"""
+
+from typing import Annotated
+
+from fastapi import APIRouter, Query
+
+from llama_stack_api.router_utils import PUBLIC_ROUTE_KEY, standard_responses
+from llama_stack_api.version import LLAMA_STACK_API_V1
+
+from .api import Inspect
+from .models import (
+    ApiFilter,
+    HealthInfo,
+    ListRoutesResponse,
+    VersionInfo,
+)
+
+
+def create_router(impl: Inspect) -> APIRouter:
+    """Create a FastAPI router for the Inspect API."""
+    router = APIRouter(
+        prefix=f"/{LLAMA_STACK_API_V1}",
+        tags=["Inspect"],
+        responses=standard_responses,
+    )
+
+    @router.get(
+        "/inspect/routes",
+        response_model=ListRoutesResponse,
+        summary="List routes.",
+        description="List all available API routes with their methods and implementing providers.",
+        responses={200: {"description": "Response containing information about all available routes."}},
+    )
+    async def list_routes(
+        api_filter: Annotated[
+            ApiFilter | None,
+            Query(
+                description="Optional filter to control which routes are returned. Can be an API level ('v1', 'v1alpha', 'v1beta') to show non-deprecated routes at that level, or 'deprecated' to show deprecated routes across all levels. If not specified, returns all non-deprecated routes."
+            ),
+        ] = None,
+    ) -> ListRoutesResponse:
+        return await impl.list_routes(api_filter)
+
+    @router.get(
+        "/health",
+        response_model=HealthInfo,
+        summary="Get health status.",
+        description="Get the current health status of the service.",
+        responses={200: {"description": "Health information indicating if the service is operational."}},
+        openapi_extra={PUBLIC_ROUTE_KEY: True},
+    )
+    async def health() -> HealthInfo:
+        return await impl.health()
+
+    @router.get(
+        "/version",
+        response_model=VersionInfo,
+        summary="Get version.",
+        description="Get the version of the service.",
+        responses={200: {"description": "Version information containing the service version number."}},
+        openapi_extra={PUBLIC_ROUTE_KEY: True},
+    )
+    async def version() -> VersionInfo:
+        return await impl.version()
+
+    return router

llama_stack_api/inspect_api/models.py

@@ -0,0 +1,28 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+"""Pydantic models for Inspect API requests and responses.
+
+This module re-exports models from llama_stack_api.admin.models to ensure
+a single source of truth and avoid type conflicts.
+"""
+
+# Import and re-export shared models from admin
+from llama_stack_api.admin.models import (
+    ApiFilter,
+    HealthInfo,
+    ListRoutesResponse,
+    RouteInfo,
+    VersionInfo,
+)
+
+__all__ = [
+    "ApiFilter",
+    "RouteInfo",
+    "HealthInfo",
+    "VersionInfo",
+    "ListRoutesResponse",
+]

{llama_stack/apis/agents → llama_stack_api/internal}/__init__.py

@@ -4,4 +4,6 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
-from .agents import *
+# Internal subpackage for shared interfaces that are not part of the public API.
+
+__all__: list[str] = []

llama_stack/providers/utils/kvstore/api.py → llama_stack_api/internal/kvstore.py

@@ -9,6 +9,8 @@ from typing import Protocol
 
 
 class KVStore(Protocol):
+    """Protocol for simple key/value storage backends."""
+
     # TODO: make the value type bytes instead of str
     async def set(self, key: str, value: str, expiration: datetime | None = None) -> None: ...
 
@@ -19,3 +21,6 @@ class KVStore(Protocol):
     async def values_in_range(self, start_key: str, end_key: str) -> list[str]: ...
 
     async def keys_in_range(self, start_key: str, end_key: str) -> list[str]: ...
+
+
+__all__ = ["KVStore"]

llama_stack_api/internal/sqlstore.py

@@ -0,0 +1,79 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from collections.abc import Mapping, Sequence
+from enum import Enum
+from typing import Any, Literal, Protocol
+
+from pydantic import BaseModel
+
+from llama_stack_api import PaginatedResponse
+
+
+class ColumnType(Enum):
+    INTEGER = "INTEGER"
+    STRING = "STRING"
+    TEXT = "TEXT"
+    FLOAT = "FLOAT"
+    BOOLEAN = "BOOLEAN"
+    JSON = "JSON"
+    DATETIME = "DATETIME"
+
+
+class ColumnDefinition(BaseModel):
+    type: ColumnType
+    primary_key: bool = False
+    nullable: bool = True
+    default: Any = None
+
+
+class SqlStore(Protocol):
+    """Protocol for common SQL-store functionality."""
+
+    async def create_table(self, table: str, schema: Mapping[str, ColumnType | ColumnDefinition]) -> None: ...
+
+    async def insert(self, table: str, data: Mapping[str, Any] | Sequence[Mapping[str, Any]]) -> None: ...
+
+    async def upsert(
+        self,
+        table: str,
+        data: Mapping[str, Any],
+        conflict_columns: list[str],
+        update_columns: list[str] | None = None,
+    ) -> None: ...
+
+    async def fetch_all(
+        self,
+        table: str,
+        where: Mapping[str, Any] | None = None,
+        where_sql: str | None = None,
+        limit: int | None = None,
+        order_by: list[tuple[str, Literal["asc", "desc"]]] | None = None,
+        cursor: tuple[str, str] | None = None,
+    ) -> PaginatedResponse: ...
+
+    async def fetch_one(
+        self,
+        table: str,
+        where: Mapping[str, Any] | None = None,
+        where_sql: str | None = None,
+        order_by: list[tuple[str, Literal["asc", "desc"]]] | None = None,
+    ) -> dict[str, Any] | None: ...
+
+    async def update(self, table: str, data: Mapping[str, Any], where: Mapping[str, Any]) -> None: ...
+
+    async def delete(self, table: str, where: Mapping[str, Any]) -> None: ...
+
+    async def add_column_if_not_exists(
+        self,
+        table: str,
+        column_name: str,
+        column_type: ColumnType,
+        nullable: bool = True,
+    ) -> None: ...
+
+
+__all__ = ["ColumnDefinition", "ColumnType", "SqlStore"]

{llama_stack/apis/models → llama_stack_api}/models.py

@@ -9,10 +9,9 @@ from typing import Any, Literal, Protocol, runtime_checkable
 
 from pydantic import BaseModel, ConfigDict, Field, field_validator
 
-from llama_stack.apis.resource import Resource, ResourceType
-from llama_stack.apis.version import LLAMA_STACK_API_V1
-from llama_stack.providers.utils.telemetry.trace_protocol import trace_protocol
-from llama_stack.schema_utils import json_schema_type, webmethod
+from llama_stack_api.resource import Resource, ResourceType
+from llama_stack_api.schema_utils import json_schema_type, webmethod
+from llama_stack_api.version import LLAMA_STACK_API_V1
 
 
 class CommonModelFields(BaseModel):
@@ -27,10 +26,12 @@ class ModelType(StrEnum):
     """Enumeration of supported model types in Llama Stack.
     :cvar llm: Large language model for text generation and completion
     :cvar embedding: Embedding model for converting text to vector representations
+    :cvar rerank: Reranking model for reordering documents based on their relevance to a query
     """
 
     llm = "llm"
     embedding = "embedding"
+    rerank = "rerank"
 
 
 @json_schema_type
@@ -88,22 +89,23 @@ class OpenAIModel(BaseModel):
     :object: The object type, which will be "model"
     :created: The Unix timestamp in seconds when the model was created
     :owned_by: The owner of the model
+    :custom_metadata: Llama Stack-specific metadata including model_type, provider info, and additional metadata
     """
 
     id: str
     object: Literal["model"] = "model"
     created: int
     owned_by: str
+    custom_metadata: dict[str, Any] | None = None
 
 
+@json_schema_type
 class OpenAIListModelsResponse(BaseModel):
     data: list[OpenAIModel]
 
 
 @runtime_checkable
-@trace_protocol
 class Models(Protocol):
-    @webmethod(route="/models", method="GET", level=LLAMA_STACK_API_V1)
     async def list_models(self) -> ListModelsResponse:
         """List all models.
 
@@ -111,7 +113,7 @@ class Models(Protocol):
         """
         ...
 
-    @webmethod(route="/openai/v1/models", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
+    @webmethod(route="/models", method="GET", level=LLAMA_STACK_API_V1)
     async def openai_list_models(self) -> OpenAIListModelsResponse:
         """List models using the OpenAI API.
 
@@ -133,7 +135,7 @@ class Models(Protocol):
         """
         ...
 
-    @webmethod(route="/models", method="POST", level=LLAMA_STACK_API_V1)
+    @webmethod(route="/models", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
     async def register_model(
         self,
         model_id: str,
@@ -155,7 +157,7 @@ class Models(Protocol):
         """
         ...
 
-    @webmethod(route="/models/{model_id:path}", method="DELETE", level=LLAMA_STACK_API_V1)
+    @webmethod(route="/models/{model_id:path}", method="DELETE", level=LLAMA_STACK_API_V1, deprecated=True)
     async def unregister_model(
         self,
         model_id: str,

{llama_stack/apis/agents → llama_stack_api}/openai_responses.py

@@ -4,13 +4,16 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
+from collections.abc import Sequence
+from enum import Enum
 from typing import Annotated, Any, Literal
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, model_validator
 from typing_extensions import TypedDict
 
-from llama_stack.apis.vector_io import SearchRankingOptions as FileSearchRankingOptions
-from llama_stack.schema_utils import json_schema_type, register_schema
+from llama_stack_api.inference import OpenAITokenLogProb
+from llama_stack_api.schema_utils import json_schema_type, register_schema
+from llama_stack_api.vector_io import SearchRankingOptions as FileSearchRankingOptions
 
 # NOTE(ashwin): this file is literally a copy of the OpenAI responses API schema. We should probably
 # take their YAML and generate this file automatically. Their YAML is available.
@@ -46,23 +49,66 @@ class OpenAIResponseInputMessageContentImage(BaseModel):
 
     :param detail: Level of detail for image processing, can be "low", "high", or "auto"
     :param type: Content type identifier, always "input_image"
+    :param file_id: (Optional) The ID of the file to be sent to the model.
     :param image_url: (Optional) URL of the image content
     """
 
     detail: Literal["low"] | Literal["high"] | Literal["auto"] = "auto"
     type: Literal["input_image"] = "input_image"
-    # TODO: handle file_id
+    file_id: str | None = None
     image_url: str | None = None
 
 
-# TODO: handle file content types
+@json_schema_type
+class OpenAIResponseInputMessageContentFile(BaseModel):
+    """File content for input messages in OpenAI response format.
+
+    :param type: The type of the input item. Always `input_file`.
+    :param file_data: The data of the file to be sent to the model.
+    :param file_id: (Optional) The ID of the file to be sent to the model.
+    :param file_url: The URL of the file to be sent to the model.
+    :param filename: The name of the file to be sent to the model.
+    """
+
+    type: Literal["input_file"] = "input_file"
+    file_data: str | None = None
+    file_id: str | None = None
+    file_url: str | None = None
+    filename: str | None = None
+
+    @model_validator(mode="after")
+    def validate_file_source(self) -> "OpenAIResponseInputMessageContentFile":
+        if not any([self.file_data, self.file_id, self.file_url, self.filename]):
+            raise ValueError(
+                "At least one of 'file_data', 'file_id', 'file_url', or 'filename' must be provided for file content"
+            )
+        return self
+
+
 OpenAIResponseInputMessageContent = Annotated[
-    OpenAIResponseInputMessageContentText | OpenAIResponseInputMessageContentImage,
+    OpenAIResponseInputMessageContentText
+    | OpenAIResponseInputMessageContentImage
+    | OpenAIResponseInputMessageContentFile,
     Field(discriminator="type"),
 ]
 register_schema(OpenAIResponseInputMessageContent, name="OpenAIResponseInputMessageContent")
 
 
+@json_schema_type
+class OpenAIResponsePrompt(BaseModel):
+    """OpenAI compatible Prompt object that is used in OpenAI responses.
+
+    :param id: Unique identifier of the prompt template
+    :param variables: Dictionary of variable names to OpenAIResponseInputMessageContent structure for template substitution. The substitution values can either be strings, or other Response input types
+    like images or files.
+    :param version: Version number of the prompt to use (defaults to latest if not specified)
+    """
+
+    id: str
+    variables: dict[str, OpenAIResponseInputMessageContent] | None = None
+    version: str | None = None
+
+
 @json_schema_type
 class OpenAIResponseAnnotationFileCitation(BaseModel):
     """File citation annotation for referencing specific files in response content.
@@ -129,6 +175,7 @@ class OpenAIResponseOutputMessageContentOutputText(BaseModel):
     text: str
     type: Literal["output_text"] = "output_text"
     annotations: list[OpenAIResponseAnnotations] = Field(default_factory=list)
+    logprobs: list[OpenAITokenLogProb] | None = None
 
 
 @json_schema_type
@@ -159,7 +206,7 @@ class OpenAIResponseMessage(BaseModel):
     scenarios.
     """
 
-    content: str | list[OpenAIResponseInputMessageContent] | list[OpenAIResponseOutputMessageContent]
+    content: str | Sequence[OpenAIResponseInputMessageContent] | Sequence[OpenAIResponseOutputMessageContent]
     role: Literal["system"] | Literal["developer"] | Literal["user"] | Literal["assistant"]
     type: Literal["message"] = "message"
 
@@ -211,10 +258,10 @@ class OpenAIResponseOutputMessageFileSearchToolCall(BaseModel):
     """
 
     id: str
-    queries: list[str]
+    queries: Sequence[str]
     status: str
     type: Literal["file_search_call"] = "file_search_call"
-    results: list[OpenAIResponseOutputMessageFileSearchToolCallResults] | None = None
+    results: Sequence[OpenAIResponseOutputMessageFileSearchToolCallResults] | None = None
 
 
 @json_schema_type
@@ -359,7 +406,7 @@ class OpenAIResponseText(BaseModel):
 
 
 # Must match type Literals of OpenAIResponseInputToolWebSearch below
-WebSearchToolTypes = ["web_search", "web_search_preview", "web_search_preview_2025_03_11"]
+WebSearchToolTypes = ["web_search", "web_search_preview", "web_search_preview_2025_03_11", "web_search_2025_08_26"]
 
 
 @json_schema_type
@@ -371,9 +418,12 @@ class OpenAIResponseInputToolWebSearch(BaseModel):
     """
 
     # Must match values of WebSearchToolTypes above
-    type: Literal["web_search"] | Literal["web_search_preview"] | Literal["web_search_preview_2025_03_11"] = (
-        "web_search"
-    )
+    type: (
+        Literal["web_search"]
+        | Literal["web_search_preview"]
+        | Literal["web_search_preview_2025_03_11"]
+        | Literal["web_search_2025_08_26"]
+    ) = "web_search"
     # TODO: actually use search_context_size somewhere...
     search_context_size: str | None = Field(default="medium", pattern="^low|medium|high$")
     # TODO: add user_location
@@ -443,6 +493,7 @@ class OpenAIResponseInputToolMCP(BaseModel):
     :param server_label: Label to identify this MCP server
     :param server_url: URL endpoint of the MCP server
     :param headers: (Optional) HTTP headers to include when connecting to the server
+    :param authorization: (Optional) OAuth access token for authenticating with the MCP server
     :param require_approval: Approval requirement for tool calls ("always", "never", or filter)
     :param allowed_tools: (Optional) Restriction on which tools can be used from this server
     """
@@ -451,6 +502,7 @@ class OpenAIResponseInputToolMCP(BaseModel):
     server_label: str
     server_url: str
     headers: dict[str, Any] | None = None
+    authorization: str | None = Field(default=None, exclude=True)
 
     require_approval: Literal["always"] | Literal["never"] | ApprovalFilter = "never"
     allowed_tools: list[str] | AllowedToolsFilter | None = None
@@ -490,6 +542,105 @@ OpenAIResponseTool = Annotated[
 register_schema(OpenAIResponseTool, name="OpenAIResponseTool")
 
 
+@json_schema_type
+class OpenAIResponseInputToolChoiceAllowedTools(BaseModel):
+    """Constrains the tools available to the model to a pre-defined set.
+
+    :param mode: Constrains the tools available to the model to a pre-defined set
+    :param tools: A list of tool definitions that the model should be allowed to call
+    :param type: Tool choice type identifier, always "allowed_tools"
+    """
+
+    mode: Literal["auto", "required"] = "auto"
+    tools: list[dict[str, str]]
+    type: Literal["allowed_tools"] = "allowed_tools"
+
+
+@json_schema_type
+class OpenAIResponseInputToolChoiceFileSearch(BaseModel):
+    """Indicates that the model should use file search to generate a response.
+
+    :param type: Tool choice type identifier, always "file_search"
+    """
+
+    type: Literal["file_search"] = "file_search"
+
+
+@json_schema_type
+class OpenAIResponseInputToolChoiceWebSearch(BaseModel):
+    """Indicates that the model should use web search to generate a response
+
+    :param type: Web search tool type variant to use
+    """
+
+    type: (
+        Literal["web_search"]
+        | Literal["web_search_preview"]
+        | Literal["web_search_preview_2025_03_11"]
+        | Literal["web_search_2025_08_26"]
+    ) = "web_search"
+
+
+@json_schema_type
+class OpenAIResponseInputToolChoiceFunctionTool(BaseModel):
+    """Forces the model to call a specific function.
+
+    :param name: The name of the function to call
+    :param type: Tool choice type identifier, always "function"
+    """
+
+    name: str
+    type: Literal["function"] = "function"
+
+
+@json_schema_type
+class OpenAIResponseInputToolChoiceMCPTool(BaseModel):
+    """Forces the model to call a specific tool on a remote MCP server
+
+    :param server_label: The label of the MCP server to use.
+    :param type: Tool choice type identifier, always "mcp"
+    :param name: (Optional) The name of the tool to call on the server.
+    """
+
+    server_label: str
+    type: Literal["mcp"] = "mcp"
+    name: str | None = None
+
+
+@json_schema_type
+class OpenAIResponseInputToolChoiceCustomTool(BaseModel):
+    """Forces the model to call a custom tool.
+
+    :param type: Tool choice type identifier, always "custom"
+    :param name: The name of the custom tool to call.
+    """
+
+    type: Literal["custom"] = "custom"
+    name: str
+
+
+class OpenAIResponseInputToolChoiceMode(str, Enum):
+    auto = "auto"
+    required = "required"
+    none = "none"
+
+
+OpenAIResponseInputToolChoiceObject = Annotated[
+    OpenAIResponseInputToolChoiceAllowedTools
+    | OpenAIResponseInputToolChoiceFileSearch
+    | OpenAIResponseInputToolChoiceWebSearch
+    | OpenAIResponseInputToolChoiceFunctionTool
+    | OpenAIResponseInputToolChoiceMCPTool
+    | OpenAIResponseInputToolChoiceCustomTool,
+    Field(discriminator="type"),
+]
+
+# 3. Final Union without registration or None (Keep it clean)
+OpenAIResponseInputToolChoice = OpenAIResponseInputToolChoiceMode | OpenAIResponseInputToolChoiceObject
+
+register_schema(OpenAIResponseInputToolChoice, name="OpenAIResponseInputToolChoice")
+
+
 class OpenAIResponseUsageOutputTokensDetails(BaseModel):
     """Token details for output tokens in OpenAI response usage.
 
@@ -536,16 +687,20 @@ class OpenAIResponseObject(BaseModel):
     :param model: Model identifier used for generation
     :param object: Object type identifier, always "response"
     :param output: List of generated output items (messages, tool calls, etc.)
-    :param parallel_tool_calls: Whether tool calls can be executed in parallel
+    :param parallel_tool_calls: (Optional) Whether to allow more than one function tool call generated per turn.
     :param previous_response_id: (Optional) ID of the previous response in a conversation
+    :param prompt: (Optional) Reference to a prompt template and its variables.
     :param status: Current status of the response generation
     :param temperature: (Optional) Sampling temperature used for generation
     :param text: Text formatting configuration for the response
     :param top_p: (Optional) Nucleus sampling parameter used for generation
     :param tools: (Optional) An array of tools the model may call while generating a response.
+    :param tool_choice: (Optional) Tool choice configuration for the response.
     :param truncation: (Optional) Truncation strategy applied to the response
     :param usage: (Optional) Token usage information for the response
     :param instructions: (Optional) System message inserted into the model's context
+    :param max_tool_calls: (Optional) Max number of total calls to built-in tools that can be processed in a response
+    :param metadata: (Optional) Dictionary of metadata key-value pairs
     """
 
     created_at: int
@@ -553,19 +708,23 @@ class OpenAIResponseObject(BaseModel):
     id: str
     model: str
     object: Literal["response"] = "response"
-    output: list[OpenAIResponseOutput]
-    parallel_tool_calls: bool = False
+    output: Sequence[OpenAIResponseOutput]
+    parallel_tool_calls: bool | None = True
     previous_response_id: str | None = None
+    prompt: OpenAIResponsePrompt | None = None
     status: str
     temperature: float | None = None
     # Default to text format to avoid breaking the loading of old responses
     # before the field was added. New responses will have this set always.
     text: OpenAIResponseText = OpenAIResponseText(format=OpenAIResponseTextFormat(type="text"))
     top_p: float | None = None
-    tools: list[OpenAIResponseTool] | None = None
+    tools: Sequence[OpenAIResponseTool] | None = None
+    tool_choice: OpenAIResponseInputToolChoice | None = None
     truncation: str | None = None
     usage: OpenAIResponseUsage | None = None
     instructions: str | None = None
+    max_tool_calls: int | None = None
+    metadata: dict[str, str] | None = None
 
 
 @json_schema_type
@@ -691,6 +850,7 @@ class OpenAIResponseObjectStreamResponseOutputTextDelta(BaseModel):
     :param content_index: Index position within the text content
     :param delta: Incremental text content being added
     :param item_id: Unique identifier of the output item being updated
+    :param logprobs: (Optional) Token log probability details
     :param output_index: Index position of the item in the output list
     :param sequence_number: Sequential number for ordering streaming events
     :param type: Event type identifier, always "response.output_text.delta"
@@ -699,6 +859,7 @@ class OpenAIResponseObjectStreamResponseOutputTextDelta(BaseModel):
     content_index: int
     delta: str
     item_id: str
+    logprobs: list[OpenAITokenLogProb] | None = None
     output_index: int
     sequence_number: int
     type: Literal["response.output_text.delta"] = "response.output_text.delta"
@@ -889,7 +1050,7 @@ class OpenAIResponseContentPartOutputText(BaseModel):
     type: Literal["output_text"] = "output_text"
     text: str
     annotations: list[OpenAIResponseAnnotations] = Field(default_factory=list)
-    logprobs: list[dict[str, Any]] | None = None
+    logprobs: list[OpenAITokenLogProb] | None = None
 
 
 @json_schema_type
@@ -1254,20 +1415,16 @@ class OpenAIResponseInputFunctionToolCallOutput(BaseModel):
 
 OpenAIResponseInput = Annotated[
     # Responses API allows output messages to be passed in as input
-    OpenAIResponseOutputMessageWebSearchToolCall
-    | OpenAIResponseOutputMessageFileSearchToolCall
-    | OpenAIResponseOutputMessageFunctionToolCall
+    OpenAIResponseOutput
     | OpenAIResponseInputFunctionToolCallOutput
-    | OpenAIResponseMCPApprovalRequest
     | OpenAIResponseMCPApprovalResponse
-    | OpenAIResponseOutputMessageMCPCall
-    | OpenAIResponseOutputMessageMCPListTools
     | OpenAIResponseMessage,
     Field(union_mode="left_to_right"),
 ]
 register_schema(OpenAIResponseInput, name="OpenAIResponseInput")
 
 
+@json_schema_type
 class ListOpenAIResponseInputItem(BaseModel):
     """List container for OpenAI response input items.
 
@@ -1275,7 +1432,7 @@ class ListOpenAIResponseInputItem(BaseModel):
     :param object: Object type identifier, always "list"
     """
 
-    data: list[OpenAIResponseInput]
+    data: Sequence[OpenAIResponseInput]
     object: Literal["list"] = "list"
 
 
@@ -1286,7 +1443,7 @@ class OpenAIResponseObjectWithInput(OpenAIResponseObject):
     :param input: List of input items that led to this response
     """
 
-    input: list[OpenAIResponseInput]
+    input: Sequence[OpenAIResponseInput]
 
     def to_response_object(self) -> OpenAIResponseObject:
         """Convert to OpenAIResponseObject by excluding input field."""
@@ -1304,7 +1461,7 @@ class ListOpenAIResponseObject(BaseModel):
     :param object: Object type identifier, always "list"
     """
 
-    data: list[OpenAIResponseObjectWithInput]
+    data: Sequence[OpenAIResponseObjectWithInput]
     has_more: bool
     first_id: str
     last_id: str