Create policy | Lakera API documentation

curl -X POST https://platform.lakera.ai/api/v1-beta/policies \
     -H "Authorization: Bearer <token>" \
     -H "Content-Type: application/json" \
     -d '{
  "name": "string",
  "policy_mode": "IO",
  "input_detectors": [
    {
      "type": "prompt_attack",
      "threshold": "l1_confident"
    }
  ],
  "output_detectors": [
    {
      "type": "prompt_attack",
      "threshold": "l1_confident"
    }
  ]
}'

{
  "status": "ok",
  "message": "string",
  "id": "string",
  "data": {
    "name": "string",
    "policy_mode": "IO",
    "input_detectors": [
      {
        "type": "prompt_attack",
        "threshold": "l1_confident",
        "allowed_domains": [
          "my-trusted-domain.com",
          "another-safe-domain.com"
        ]
      }
    ],
    "output_detectors": [
      {
        "type": "prompt_attack",
        "threshold": "l1_confident",
        "allowed_domains": [
          "my-trusted-domain.com",
          "another-safe-domain.com"
        ]
      }
    ]
  }
}

Create a new policy configuration

Authentication

AuthorizationBearer

Bearer authentication using API key. Generate an API key from the Lakera platform. Example: Bearer sk_123…

Request

This endpoint expects an object.

namestringRequired

A descriptive name for the policy

policy_modeenumRequired

Set to "IO" for screening model inputs and outputs

Allowed values:

input_detectorslist of objectsRequired

List of detector configurations to apply to LLM inputs

output_detectorslist of objectsRequired

List of detector configurations to apply to LLM outputs

Response

Policy created successfully

statusenum

Response status

Allowed values:

messagestring

Human-readable response message

idstring

Policy ID

dataobject

Policy details

$	curl -X POST https://platform.lakera.ai/api/v1-beta/policies \
>	-H "Authorization: Bearer <token>" \
>	-H "Content-Type: application/json" \
>	-d '{
>	"name": "string",
>	"policy_mode": "IO",
>	"input_detectors": [
>	{
>	"type": "prompt_attack",
>	"threshold": "l1_confident"
>	}
>	],
>	"output_detectors": [
>	{
>	"type": "prompt_attack",
>	"threshold": "l1_confident"
>	}
>	]
>	}'

1	{
2	"status": "ok",
3	"message": "string",
4	"id": "string",
5	"data": {
6	"name": "string",
7	"policy_mode": "IO",
8	"input_detectors": [
9	{
10	"type": "prompt_attack",
11	"threshold": "l1_confident",
12	"allowed_domains": [
13	"my-trusted-domain.com",
14	"another-safe-domain.com"
15	]
16	}
17	],
18	"output_detectors": [
19	{
20	"type": "prompt_attack",
21	"threshold": "l1_confident",
22	"allowed_domains": [
23	"my-trusted-domain.com",
24	"another-safe-domain.com"
25	]
26	}
27	]
28	}
29	}