GET

/v1/schemas/moonshot/chat/v1/responses

Accept:

kind inputactivity chathash a862a5d2…extracted Oct 24, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `f6c78afc0c20eb8d222859e3bae823a1c32bb4f8876059aa61df062c264a7fdf` { "fetchedAt": 1792819279, "extractorVersion": "1", "verifiedAt": null }

Request body

13 properties · 2 required · 16 $defs

property	type	description	constraints
model*	string	ID of the model to use. This endpoint currently supports `kimi-k3.1-pro`, `kimi-k3.1-pro-highspeed`, and `kimi-k3`.
input*	string \| ResponsesInputItem[]	Input for this request. A string is equivalent to a single user message. An array holds ordered typed items and may contain conversation history, tool calls, and tool results.
instructions	string	Top-level system instructions, applied ahead of every input item.
stream	boolean	When true, the response is delivered as a stream of SSE events.	default: false
max_output_tokens	integer	Maximum number of tokens to generate for this response. For Kimi K3-series models (`kimi-k3.1-pro`, `kimi-k3.1-pro-highspeed`, and `kimi-k3`) it defaults to 131072 and can be set up to 1048576. This refers to the length of tokens you expect us to return, not t…
reasoning	object	Reasoning configuration.
text	object	Output text configuration.
tools	ResponsesTool[]	List of tools the model may call.
tool_choice	enum	Controls tool-calling behavior. With `auto`, the model decides whether to call a tool.	auto
include	enum[]	Additional fields to return. Only effective when the `web_search` tool is used. `web_search_call.action.sources` returns the web pages the search hit; `web_search_call.results` returns image search results.
prompt_cache_key	string	Context cache identifier. Reusing the same value across a session improves cache hit rate.
prompt_cache_options	object	Context cache write options. When omitted, cache write is enabled by default (5m tier): the system automatically writes the request prefix to the 5m cache tier
safety_identifier	string	A stable identifier used to help detect users of your application that may be violating usage policies. The ID should be a string that uniquely identifies each user. It is recommended to hash the username or email address to avoid sending any identifying infor…

Types

16 $defs

ResponsesAdditionalToolsItemobject

Adds callable tools partway through a conversation, effective from the position of this item.

type*	enum	additional_tools
id	string
role*	enum	developer
tools*	ResponsesTool[]

ResponsesCustomToolobject

A custom tool that takes free-form text input. Only the custom tool named `apply_patch` is supported, and it must define a `grammar` + `lark` format; when the model calls it, a `custom_tool_call` item is returned in the output.

type*	enum		custom
name*	enum	Tool name. Only `apply_patch` is supported.	apply_patch
description	string
format*	object	Input format constraint. Only the `grammar` type with `lark` syntax is supported.

ResponsesCustomToolCallItemobject

Replays a custom tool call.

type*	enum		custom_tool_call
id	string
call_id*	string	Call ID paired with the matching `custom_tool_call_output`.
name*	string
namespace	string	Namespace the tool belongs to, returned for namespace tool calls.
input*	string	Free-form text input generated by the model.
status	enum		completed

ResponsesCustomToolCallOutputItemobject

Result of executing a custom tool call.

type*	enum		custom_tool_call_output
call_id*	string	Same call ID as the matching `custom_tool_call`.
output*	string \| ResponsesInputContentPart[]	Content returned by the tool.
status	enum		completed

ResponsesFunctionCallItemobject

Replays a function call.

type*	enum		function_call
id	string
call_id*	string	Call ID paired with the matching `function_call_output`.
name*	string
namespace	string	Namespace the tool belongs to, returned for namespace tool calls.
arguments*	string	JSON string of the function arguments.
status	enum		completed

ResponsesFunctionCallOutputItemobject

Result of executing a function call.

type*	enum		function_call_output
call_id*	string	Same call ID as the matching `function_call`.
output*	string \| ResponsesInputContentPart[]	Content returned by the tool.
status	enum		completed

ResponsesFunctionToolobject

A function tool whose arguments are described by a JSON Schema.

type*	enum		function
name*	string	Function name. Must follow the regex: ^[a-zA-Z_][a-zA-Z0-9-_]{0,127}$
description	string
parameters	object	JSON Schema describing the function arguments.
strict	boolean

ResponsesInputContentPartInput text | Input image | Output text

An element of a content array, discriminated by `type`.

ResponsesInputItemMessage | Reasoning | Function call | …

An element of the input array, discriminated by `type`. When `type` is omitted the item is treated as `message`.

ResponsesMessageItemobject

A conversation message. `type` may be omitted.

type	enum		message
role*	enum	Message role. `developer` is handled as a system instruction.	user \| assistant \| developer
content*	string \| ResponsesInputContentPart[]	Message content, either a string or an array of content parts.
status	enum		completed

ResponsesNamespaceToolobject

Groups a set of function or custom tools under one namespace.

type*	enum	namespace
name*	string
description*	string
tools*	ResponsesFunctionTool \| ResponsesCustomTool[]

ResponsesReasoningItemobject

Replays reasoning content from a previous turn. `content` takes precedence over `summary`.

type*	enum	reasoning
id	string
summary	object[]
content	object[]
status	enum	completed

ResponsesToolFunction tool | Custom tool | Namespace tool | …

A tool definition, discriminated by `type`. Supported types are `function`, `custom` (only `apply_patch`), `namespace`, and `web_search`; other tool types are not supported.

ResponsesToolChoiceenum

Controls tool-calling behavior. With `auto`, the model decides whether to call a tool.

auto

ResponsesWebSearchCallItemobject

Replays a previous web search call. Kept for history only and ignored during conversion; the search results are already carried by the assistant message that follows.

type*	enum	web_search_call
id	string
status	enum	completed
action	object

ResponsesWebSearchToolobject

A web search tool executed on the server side. When present, the server first decides from the input whether a search is needed; if so, it runs the search, injects the results into the model context, and returns a `web_search_call` item in the output. At most …

type*	enum		web_search
filters	object	Search scope filters.
search_content_types	enum[]	Result types to search for. Defaults to `text` only. Including `image` additionally runs an image search.	default: ["text"]
image_settings	object	Image search settings. Only effective when `search_content_types` includes `image`.

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"moonshot","endpointId":"v1/responses","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}