GET

/v1/schemas/nvidia/chat/meta/llama-guard-4-12b?kind=output

Accept:

kind outputactivity chathash c7dac4c4…extracted Oct 9, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `876b01a02db8cf38d07edb6020fc7e7f2ce1d437fbcba3c82b515cdd1e54ddea` { "fetchedAt": 1791535404, "extractorVersion": "1", "verifiedAt": null }

Response body

6 properties · 3 required · 3 $defs

property	type	constraints
model*	string
choices*	ChatCompletionResponseChoice[]
usage*	UsageInfo
id	string
object	enum	chat.completion · default: "chat.completion"
created	integer

Types

3 $defs

ChatCompletionResponseChoiceobject

index*	integer
message*	ChatMessage
finish_reason	enum	stop \| length · nullable
stop_reason	integer \| string	nullable

ChatMessageobject

role*	string
content	string		nullable

UsageInfoobject

prompt_tokens	integer	default: 0
total_tokens	integer	default: 0
completion_tokens	integer	nullable

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"nvidia","endpointId":"meta/llama-guard-4-12b","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}