GET

/v1/schemas/nvidia/chat/fastino/gliner2-pii

kind inputactivity chathash dd441062…extracted Nov 20, 2026version currentsource upstream spec ↗
fieldrecorded source
schemaSource linked above.

upstream-spec

Source trace
372cfa2cbef5cb0898a18ddf3f076a87dda2f37478a6296976263ea43254a8c3
{
  "fetchedAt": 1795152072,
  "extractorVersion": "2",
  "verifiedAt": null
}

Request body

9 properties · 3 required · 2 $defs
propertytypedescriptionconstraints
model*stringlength ≥ 1 · length ≤ 200 · default: "fastino/gliner2-pii"
messages*Message[]Only the last message is analyzed. Text parts are concatenated without a separator. Maximum analyzed text is 16384 characters and 128 chunks.items ≥ 1 · items ≤ 32
labels*enum[]Required nonempty subset of the 42 supported PII labels; duplicates are rejected.items ≥ 1 · items ≤ 42
thresholdnumberGlobal confidence cutoff. This is an extraction score, not a guarantee of PII detection.0.01 ≤ n ≤ 1 · default: 0.3
chunk_lengthintegerChunk length in Unicode characters; must exceed overlap.64 ≤ n ≤ 2048 · default: 384
overlapintegerOverlap in Unicode characters, not tokens or words.0 ≤ n ≤ 512 · default: 128
flat_nerbooleandefault: false
batch_sizeinteger1 ≤ n ≤ 16 · default: 8
streamconstfalse · default: false

Types

2 $defs

Messageobject

role*enumuser | assistant | system
content*string | TextPart[]

TextPartobject

type*const"text"
text*stringlength ≤ 16384

Validate a payload

before you spend tokens
shell
$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"nvidia","endpointId":"fastino/gliner2-pii","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}