GET

/v1/schemas/anthropic/chat/v1/messages?kind=output

Accept:

kind outputactivity chathash 7764f96a…extracted Nov 11, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `690622377a5eedc0e9f2265ed755ee34dc0c52d35b47327f425777a550f45e1e` { "fetchedAt": 1794374425, "extractorVersion": "2", "verifiedAt": null }

Response body

11 properties · 11 required · 68 $defs

property	type	description	constraints
id*	string	Unique object identifier.
type*	const	Object type.	"message" · default: "message"
role*	const	Conversational role of the generated message.	"assistant" · default: "assistant"
content*	ContentBlock[]	Content generated by the model.
model*	string	The model that will complete your prompt.	22 known values
stop_reason*	enum	The reason that we stopped. This may be one the following values: * `"end_turn"`: the model reached a natural stopping point * `"max_tokens"`: we exceeded the requested `max_tokens` or the model's maximum * `"stop_sequence"`: one of your provided custom `stop_sequences` was generated * `"tool_use"`: the model invoked one or more tools * `"pause_turn"`: we paused a long-running turn. You may provide the response back as-is in a subsequent request to let the model continue. * `"refusal"`: when streaming classifiers intervene to handle potential policy violations * `"model_context_window_exceeded"`: we exceeded the model's context window In non-streaming mode this value is always non-null. In streaming mode, it is null in the `message_start` event and non-null otherwise.	end_turn \| max_tokens \| stop_sequence \| tool_use \| pause_turn \| refusal · +1 more · nullable
stop_sequence*	string	Which custom stop sequence was generated, if any. This value will be a non-null string if one of your custom stop sequences was generated.	nullable
stop_details*	RefusalStopDetails	Structured information about a refusal.	nullable
usage*	Usage	Billing and rate-limit usage.
diagnostics*	Diagnostics	Request-level diagnostics: why the prompt cache could not fully reuse the prefix of the request named by `diagnostics.previous_message_id`.	nullable
container*	Container	Information about the container used in the request (for the code execution tool)	nullable

Types

68 $defs

Base64PDFSourceobject

data*	string	format: byte
media_type*	const	"application/pdf"
type*	const	"base64"

BashCodeExecutionToolResultErrorCodeenum

invalid_tool_input | unavailable | too_many_requests | execution_time_exceeded | output_file_too_large

CacheCreationobject

ephemeral_1h_input_tokens*	integer	The number of input tokens used to create the 1 hour cache entry.	n ≥ 0 · default: 0
ephemeral_5m_input_tokens*	integer	The number of input tokens used to create the 5 minute cache entry.	n ≥ 0 · default: 0

CacheMissMessagesChangedobject

A `cache_miss_reason` for a request whose `messages` differ from the previous request's in an earlier entry, instead of only having entries appended. The `model`, `system`, and `tools` match.

cache_missed_input_tokens*	integer	Approximate number of input tokens that would have been read from cache had the prefix matched the previous request.
type*	const		"messages_changed" · default: "messages_changed"

CacheMissModelChangedobject

A `cache_miss_reason` for a request whose `model` or `speed` differs from the previous request's.

cache_missed_input_tokens*	integer	Approximate number of input tokens that would have been read from cache had the prefix matched the previous request.
type*	const		"model_changed" · default: "model_changed"

CacheMissPreviousMessageNotFoundobject

A `cache_miss_reason` for when the API doesn't have a stored record of the request that `diagnostics.previous_message_id` references.

type*

const

"previous_message_not_found" · default: "previous_message_not_found"

CacheMissReasonCacheMissModelChanged | CacheMissSystemChanged | CacheMissToolsChanged | …

CacheMissSystemChangedobject

A `cache_miss_reason` for a request whose `system` differs from the previous request's.

cache_missed_input_tokens*	integer	Approximate number of input tokens that would have been read from cache had the prefix matched the previous request.
type*	const		"system_changed" · default: "system_changed"

CacheMissToolsChangedobject

A `cache_miss_reason` for a request whose `tools` differ from the previous request's, for example because tools were added, removed, or reordered.

cache_missed_input_tokens*	integer	Approximate number of input tokens that would have been read from cache had the prefix matched the previous request.
type*	const		"tools_changed" · default: "tools_changed"

CacheMissUnavailableobject

A `cache_miss_reason` for when the API can't give a diagnosis. The request itself was processed normally.

type*

const

"unavailable" · default: "unavailable"

CodeExecutionToolResultErrorCodeenum

invalid_tool_input | unavailable | too_many_requests | execution_time_exceeded

Containerobject

Information about the container used in the request (for the code execution tool)

expires_at*	string	The time at which the container will expire.	format: date-time
id*	string	Identifier for the container used in this request
skills*	ContainerSkill[]	Skills loaded in the container	nullable

ContainerSkillobject

A skill that was loaded in a container (response model).

skill_id*	string	Skill ID	length ≥ 1 · length ≤ 64
type*	enum	Type of skill - either 'anthropic' (built-in) or 'custom' (user-defined)	anthropic \| custom
version*	string	The resolved version: a skill version ID for custom skills.	length ≥ 1 · length ≤ 64

ContentBlockResponseTextBlock | ResponseThinkingBlock | ResponseRedactedThinkingBlock | …

Diagnosticsobject

Request-level diagnostics: why the prompt cache could not fully reuse the prefix of the request named by `diagnostics.previous_message_id`.

cache_miss_reason*

CacheMissModelChanged | CacheMissSystemChanged | CacheMissToolsChanged | …

Explains why the prompt cache could not fully reuse the prefix from the request identified by `diagnostics.previous_message_id`. `null` means diagnosis is still pending — the response was serialized before the background comparison completed.

nullable

DirectCallerobject

Tool invocation directly from the model.

type*

const

"direct"

Modelstring

The model that will complete your prompt.

22 known values

OutputTokensDetailsobject

thinking_tokens*

integer

Number of output tokens the model generated as internal reasoning, including the thinking-block delimiter tokens.

n ≥ 0 · default: 0

PlainTextSourceobject

data*	string
media_type*	const	"text/plain"
type*	const	"text"

RefusalCategorystring

The policy category that triggered a refusal.

5 known values

RefusalStopDetailsobject

Structured information about a refusal.

category*	string	The policy category that triggered a refusal.	5 known values · nullable
explanation*	string	Human-readable explanation of the refusal. This text is not guaranteed to be stable. `null` when no explanation is available for the category.	nullable
type*	const		"refusal" · default: "refusal"

ResponseBashCodeExecutionOutputBlockobject

file_id*	string
type*	const		"bash_code_execution_output" · default: "bash_code_execution_output"

ResponseBashCodeExecutionResultBlockobject

content*	ResponseBashCodeExecutionOutputBlock[]
return_code*	integer
stderr*	string
stdout*	string
type*	const	"bash_code_execution_result" · default: "bash_code_execution_result"

ResponseBashCodeExecutionToolResultBlockobject

content*	ResponseBashCodeExecutionToolResultError \| ResponseBashCodeExecutionResultBlock
tool_use_id*	string
type*	const	"bash_code_execution_tool_result… · default: "bash_code_execution_tool_result…

ResponseBashCodeExecutionToolResultErrorobject

error_code*	enum		invalid_tool_input \| unavailable \| too_many_requests \| execution_time_exceeded \| output_file_too_large
type*	const		"bash_code_execution_tool_result… · default: "bash_code_execution_tool_result…

ResponseCharLocationCitationobject

cited_text*	string
document_index*	integer	n ≥ 0
document_title*	string	nullable
end_char_index*	integer
file_id*	string	nullable
start_char_index*	integer	n ≥ 0
type*	const	"char_location" · default: "char_location"

ResponseCitationsConfigobject

enabled*

boolean

default: false

ResponseCodeExecutionOutputBlockobject

file_id*	string
type*	const		"code_execution_output" · default: "code_execution_output"

ResponseCodeExecutionResultBlockobject

content*	ResponseCodeExecutionOutputBlock[]
return_code*	integer
stderr*	string
stdout*	string
type*	const	"code_execution_result" · default: "code_execution_result"

ResponseCodeExecutionToolResultBlockobject

content*	ResponseCodeExecutionToolResultError \| ResponseCodeExecutionResultBlock \| ResponseEncryptedCodeExecutionResultBlock
tool_use_id*	string
type*	const	"code_execution_tool_result" · default: "code_execution_tool_result"

ResponseCodeExecutionToolResultErrorobject

error_code*	enum		invalid_tool_input \| unavailable \| too_many_requests \| execution_time_exceeded
type*	const		"code_execution_tool_result_erro… · default: "code_execution_tool_result_erro…

ResponseContainerUploadBlockobject

Response model for a file uploaded to the container.

file_id*	string
type*	const		"container_upload" · default: "container_upload"

ResponseContentBlockLocationCitationobject

cited_text*	string	The full text of the cited block range, concatenated.
document_index*	integer		n ≥ 0
document_title*	string		nullable
end_block_index*	integer	Exclusive 0-based end index of the cited block range in the source's `content` array.
file_id*	string		nullable
start_block_index*	integer	0-based index of the first cited block in the source's `content` array.	n ≥ 0
type*	const		"content_block_location" · default: "content_block_location"

ResponseDocumentBlockobject

citations*	ResponseCitationsConfig	Citation configuration for the document	nullable
source*	Base64PDFSource \| PlainTextSource
title*	string	The title of the document	nullable
type*	const		"document" · default: "document"

ResponseEncryptedCodeExecutionResultBlockobject

The output and exit code of Python code that Claude ran with the code execution tool, with standard output in encrypted form.

content*	ResponseCodeExecutionOutputBlock[]
encrypted_stdout*	string
return_code*	integer
stderr*	string
type*	const	"encrypted_code_execution_result… · default: "encrypted_code_execution_result…

ResponsePageLocationCitationobject

cited_text*	string
document_index*	integer	n ≥ 0
document_title*	string	nullable
end_page_number*	integer
file_id*	string	nullable
start_page_number*	integer	n ≥ 1
type*	const	"page_location" · default: "page_location"

ResponseRedactedThinkingBlockobject

data*	string	The contents of this redacted thinking block, returned when portions of the model's thinking were safety-redacted. This field is opaque and encrypted, with no readable content.
type*	const		"redacted_thinking" · default: "redacted_thinking"

ResponseSearchResultLocationCitationobject

cited_text*	string	The full text of the cited block range, concatenated.
end_block_index*	integer	Exclusive 0-based end index of the cited block range in the source's `content` array.
search_result_index*	integer	0-based index of the cited search result among all `search_result` content blocks in the request, in the order they appear across messages and tool results.	n ≥ 0
source*	string
start_block_index*	integer	0-based index of the first cited block in the source's `content` array.	n ≥ 0
title*	string		nullable
type*	const		"search_result_location" · default: "search_result_location"

ResponseServerToolUseBlockobject

caller*	DirectCaller \| ServerToolCaller \| ServerToolCaller_20260120	default: {"type":"direct"}
id*	string
input*	object
name*	enum	web_search \| web_fetch \| code_execution \| bash_code_execution \| text_editor_code_execution \| tool_search_tool_regex · +1 more
type*	const	"server_tool_use" · default: "server_tool_use"

ResponseTextBlockobject

citations*	ResponseCharLocationCitation \| ResponsePageLocationCitation \| ResponseContentBlockLocationCitation[]	Citations supporting the text block. The type of citation returned will depend on the type of document being cited. Citing a PDF results in `page_location`, plain text results in `char_location`, and content document results in `content_block_location`.	nullable
text*	string		length ≥ 0
type*	const		"text" · default: "text"

ResponseTextEditorCodeExecutionCreateResultBlockobject

is_file_update*	boolean
type*	const		"text_editor_code_execution_crea… · default: "text_editor_code_execution_crea…

ResponseTextEditorCodeExecutionStrReplaceResultBlockobject

lines*	string[]	nullable
new_lines*	integer	nullable
new_start*	integer	nullable
old_lines*	integer	nullable
old_start*	integer	nullable
type*	const	"text_editor_code_execution_str_… · default: "text_editor_code_execution_str_…

ResponseTextEditorCodeExecutionToolResultBlockobject

content*	ResponseTextEditorCodeExecutionToolResultError \| ResponseTextEditorCodeExecutionViewResultBlock \| ResponseTextEditorCodeExecutionCreateResultBlock \| …
tool_use_id*	string
type*	const	"text_editor_code_execution_tool… · default: "text_editor_code_execution_tool…

ResponseTextEditorCodeExecutionToolResultErrorobject

error_code*	enum	invalid_tool_input \| unavailable \| too_many_requests \| execution_time_exceeded \| file_not_found
error_message*	string	nullable
type*	const	"text_editor_code_execution_tool… · default: "text_editor_code_execution_tool…

ResponseTextEditorCodeExecutionViewResultBlockobject

content*	string
file_type*	enum	text \| image \| pdf
num_lines*	integer	nullable
start_line*	integer	nullable
total_lines*	integer	nullable
type*	const	"text_editor_code_execution_view… · default: "text_editor_code_execution_view…

ResponseThinkingBlockobject

signature*	string	A value used to verify that this thinking block was generated by Claude when it is passed back to the API.
thinking*	string	The text of Claude's thinking process for this block.
type*	const		"thinking" · default: "thinking"

ResponseToolReferenceBlockobject

tool_name*	string		length ≥ 1 · length ≤ 256
type*	const		"tool_reference" · default: "tool_reference"

ResponseToolSearchToolResultBlockobject

content*	ResponseToolSearchToolResultError \| ResponseToolSearchToolSearchResultBlock
tool_use_id*	string
type*	const	"tool_search_tool_result" · default: "tool_search_tool_result"

ResponseToolSearchToolResultErrorobject

error_code*	enum	invalid_tool_input \| unavailable \| too_many_requests \| execution_time_exceeded
error_message*	string	nullable
type*	const	"tool_search_tool_result_error" · default: "tool_search_tool_result_error"

ResponseToolSearchToolSearchResultBlockobject

tool_references*	ResponseToolReferenceBlock[]
type*	const		"tool_search_tool_search_result" · default: "tool_search_tool_search_result"

ResponseToolUseBlockobject

caller*	DirectCaller \| ServerToolCaller \| ServerToolCaller_20260120		default: {"type":"direct"}
id*	string
input*	object
name*	string		length ≥ 1
toolset_name	string	For a toolset member tool_use, the toolset family.	length ≥ 1 · length ≤ 64 · nullable
type*	const		"tool_use" · default: "tool_use"

ResponseWebFetchResultBlockobject

content*	ResponseDocumentBlock
retrieved_at*	string	ISO 8601 timestamp when the content was retrieved	nullable
type*	const		"web_fetch_result" · default: "web_fetch_result"
url*	string	Fetched content URL

ResponseWebFetchToolResultBlockobject

caller*	DirectCaller \| ServerToolCaller \| ServerToolCaller_20260120	default: {"type":"direct"}
content*	ResponseWebFetchToolResultError \| ResponseWebFetchResultBlock
tool_use_id*	string
type*	const	"web_fetch_tool_result" · default: "web_fetch_tool_result"

ResponseWebFetchToolResultErrorobject

error_code*	enum		invalid_tool_input \| url_too_long \| url_not_allowed \| url_not_in_prior_context \| url_not_accessible \| unsupported_content_type · +4 more
type*	const		"web_fetch_tool_result_error" · default: "web_fetch_tool_result_error"

ResponseWebSearchResultBlockobject

encrypted_content*	string
page_age*	string	nullable
title*	string
type*	const	"web_search_result" · default: "web_search_result"
url*	string

ResponseWebSearchResultLocationCitationobject

cited_text*	string
encrypted_index*	string
title*	string	length ≤ 512 · nullable
type*	const	"web_search_result_location" · default: "web_search_result_location"
url*	string

ResponseWebSearchToolResultBlockobject

caller*	DirectCaller \| ServerToolCaller \| ServerToolCaller_20260120	default: {"type":"direct"}
content*	ResponseWebSearchToolResultError \| ResponseWebSearchResultBlock[]
tool_use_id*	string
type*	const	"web_search_tool_result" · default: "web_search_tool_result"

ResponseWebSearchToolResultErrorobject

error_code*	enum		invalid_tool_input \| unavailable \| max_uses_exceeded \| too_many_requests \| query_too_long \| request_too_large
type*	const		"web_search_tool_result_error" · default: "web_search_tool_result_error"

ServerToolCallerobject

Tool invocation generated by a server-side tool.

tool_id*	string
type*	const		"code_execution_20250825"

ServerToolCaller_20260120object

tool_id*	string
type*	const		"code_execution_20260120"

ServerToolUsageobject

web_fetch_requests*	integer	The number of web fetch tool requests.	n ≥ 0 · default: 0
web_search_requests*	integer	The number of web search tool requests.	n ≥ 0 · default: 0

Speedenum

standard | fast

StopReasonenum

TextEditorCodeExecutionToolResultErrorCodeenum

invalid_tool_input | unavailable | too_many_requests | execution_time_exceeded | file_not_found

ToolSearchToolResultErrorCodeenum

invalid_tool_input | unavailable | too_many_requests | execution_time_exceeded

Usageobject

cache_creation*	CacheCreation	Breakdown of cached tokens by TTL	nullable
cache_creation_input_tokens*	integer	The number of input tokens used to create the cache entry.	n ≥ 0 · nullable
cache_read_input_tokens*	integer	The number of input tokens read from the cache.	n ≥ 0 · nullable
inference_geo*	string	The geographic region where inference was performed for this request.	nullable
input_tokens*	integer	The number of input tokens which were used.	n ≥ 0
output_tokens*	integer	The number of output tokens which were used.	n ≥ 0
output_tokens_details*	OutputTokensDetails	Breakdown of output tokens by category. `output_tokens` remains the inclusive, authoritative total used for billing. This object provides a read-only decomposition for observability — for example, how many of the billed output tokens were spent on internal reasoning that may have been summarized before being returned to you.	nullable
server_tool_use*	ServerToolUsage	The number of server tool requests.	nullable
service_tier*	enum	If the request used the priority, standard, or batch tier.	standard \| priority \| batch · nullable
speed*	enum	The inference speed mode used for this request.	standard \| fast · nullable

WebFetchToolResultErrorCodeenum

WebSearchToolResultErrorCodeenum

Validate a payload

before you spend tokens

shell

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