gliner-multi-pii-v1 - ZeroGPU

curl --request POST \
  --url https://api.zerogpu.ai/v1/responses \
  --header 'Content-Type: application/json' \
  --header 'x-api-key: <api-key>' \
  --header 'x-project-id: <api-key>' \
  --data '
{
  "input": "Hello Jane Doe, this is John Doe reaching out regarding my recent order. If you need any additional details, feel free to call me at 415-555-0134 during business hours. You can also email me at hi@example.com, and I’ll respond as soon as possible. Looking forward to your update on the issue.",
  "model": "gliner-multi-pii-v1"
}
'

{
  "redacted_text": "Hello [PERSON], this is [PERSON] reaching out regarding my recent order. If you need any additional details, feel free to call me at [PHONE_NUMBER] during business hours. You can also email me at [EMAIL], and I will respond as soon as possible.",
  "entities": [
    {
      "text": "Jane Doe",
      "label": "person",
      "start": 6,
      "end": 14,
      "score": 0.9982
    },
    {
      "text": "John Doe",
      "label": "person",
      "start": 24,
      "end": 32,
      "score": 0.9981
    },
    {
      "text": "415-555-0134",
      "label": "phone number",
      "start": 133,
      "end": 145,
      "score": 0.9725
    },
    {
      "text": "hi@example.com",
      "label": "email",
      "start": 194,
      "end": 208,
      "score": 0.9812
    }
  ],
  "entities_by_label": {
    "person": [
      "Jane Doe",
      "John Doe"
    ],
    "phone number": [
      "415-555-0134"
    ],
    "email": [
      "hi@example.com"
    ]
  }
}

POST

/

responses

curl --request POST \
  --url https://api.zerogpu.ai/v1/responses \
  --header 'Content-Type: application/json' \
  --header 'x-api-key: <api-key>' \
  --header 'x-project-id: <api-key>' \
  --data '
{
  "input": "Hello Jane Doe, this is John Doe reaching out regarding my recent order. If you need any additional details, feel free to call me at 415-555-0134 during business hours. You can also email me at hi@example.com, and I’ll respond as soon as possible. Looking forward to your update on the issue.",
  "model": "gliner-multi-pii-v1"
}
'

{
  "redacted_text": "Hello [PERSON], this is [PERSON] reaching out regarding my recent order. If you need any additional details, feel free to call me at [PHONE_NUMBER] during business hours. You can also email me at [EMAIL], and I will respond as soon as possible.",
  "entities": [
    {
      "text": "Jane Doe",
      "label": "person",
      "start": 6,
      "end": 14,
      "score": 0.9982
    },
    {
      "text": "John Doe",
      "label": "person",
      "start": 24,
      "end": 32,
      "score": 0.9981
    },
    {
      "text": "415-555-0134",
      "label": "phone number",
      "start": 133,
      "end": 145,
      "score": 0.9725
    },
    {
      "text": "hi@example.com",
      "label": "email",
      "start": 194,
      "end": 208,
      "score": 0.9812
    }
  ],
  "entities_by_label": {
    "person": [
      "Jane Doe",
      "John Doe"
    ],
    "phone number": [
      "415-555-0134"
    ],
    "email": [
      "hi@example.com"
    ]
  }
}

Authorizations

x-api-key

string

header

required

x-project-id

string

header

required

Body

application/json

model

string

default:gliner-multi-pii-v1

required

Model identifier (fixed for this playground). Use request examples to change use cases.

Allowed value: "gliner-multi-pii-v1"

Example:

"gliner-multi-pii-v1"

input

string<textarea>

required

Multi-line text or document content to send to the model.

Required string length: 1 - 131072

metadata

object

Use-case options for the model. Two use cases are supported:

extract-pii — return detected PII entities without altering the source text.
redact — mask detected PII in-line in the returned redacted_text (e.g. [PERSON], [EMAIL]).

Hide child attributes

metadata.usecase

enum<string>

default:extract-pii

Which operation to run. extract-pii returns only the detected entities; redact also returns a masked redacted_text.

Available options:

extract-pii,

redact

metadata.mask

enum<string>

Redaction style, used when usecase is redact. label replaces each entity with a typed placeholder such as [PERSON] or [EMAIL].

Available options:

label

metadata.threshold

number

default:0.5

Minimum confidence score (0–1) an entity must reach to be returned. Lower values surface more, lower-confidence matches.

Required range: 0 <= x <= 1

metadata.categories

string[]

Restrict detection to specific PII category groups (e.g. identity, contact). Omit to detect all supported categories.

Example:

["identity", "contact"]

Response

Success

The response is of type object.