Text Content Moderation API — POST /v1/moderations

POST /v1/moderations checks text for policy violations across 13 categories including violence, harassment, hate speech, and self-harm.

Use the /v1/moderations endpoint to check text content for policy violations. The response includes a per-category boolean flag indicating whether each policy category was triggered, along with confidence scores for each category — giving you fine-grained control over how you apply moderation thresholds in your application.

Request

POST https://kiosapi.com/v1/moderations

Headers

Content-Type · string · required

Must be application/json.

Authorization · string · required

Your API key in the format Bearer sk-xxx.

Request Body

model · string · required

The moderation model to use. Example: text-moderation-latest.

input · string · required

The text content you want to check for policy violations.

Example Request

cURL

curl https://kiosapi.com/v1/moderations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxx" \
  -d '{
    "model": "text-moderation-latest",
    "input": "She is very beautiful, I like..."
  }'

Python

import openai

openai.api_key = "sk-xxx"
openai.base_url = "https://kiosapi.com/v1/"

response = openai.moderations.create(
    model="text-moderation-latest",
    input="She is very beautiful, I like..."
)

print(response.results[0].flagged)

Categories

The moderation model evaluates your input across the following 13 categories:

Category	Description
sexual	Sexually explicit content
sexual/minors	Sexual content involving minors
harassment	Harassing or bullying language
harassment/threatening	Harassment that includes threats
hate	Hate speech targeting protected groups
hate/threatening	Hate speech that includes threats of violence
illicit	Content promoting illegal activities
illicit/violent	Illegal activities involving violence
self-harm	Content depicting or promoting self-harm
self-harm/intent	Expressed intent to engage in self-harm
self-harm/instructions	Instructions for self-harm
violence	Violent content or descriptions
violence/graphic	Graphic or gratuitous violent content

The top-level flagged field is true if any of the 13 categories is triggered, providing a convenient single signal for content filtering. Use the individual categories and category_scores fields for more granular control.

Response

{
  "id": "modr-970d409ef3bef3b70c73d8232df86e7d",
  "model": "text-moderation-latest",
  "results": [
    {
      "flagged": true,
      "categories": {
        "sexual": false,
        "sexual/minors": false,
        "harassment": false,
        "harassment/threatening": false,
        "hate": false,
        "hate/threatening": false,
        "illicit": false,
        "illicit/violent": false,
        "self-harm": false,
        "self-harm/intent": false,
        "self-harm/instructions": false,
        "violence": true,
        "violence/graphic": false
      },
      "category_scores": {
        "sexual": 0.0000002341,
        "violence": 0.8599265510,
        "violence/graphic": 0.3770173639
      }
    }
  ]
}

Response Fields

id · string

A unique identifier for this moderation request.

model · string

The moderation model that processed the request.

results · array

An array of moderation result objects, one per input.

Show results[] fields

results[].flagged · boolean

true if the input triggered one or more policy categories; false otherwise. Use this as your primary pass/fail signal.

results[].categories · object

A map of each of the 13 category names to a boolean. true means that category was triggered for the input.

results[].category_scores · object

A map of category names to floating-point confidence scores. Higher values indicate higher model confidence that the category applies. You can use these scores to implement custom severity thresholds beyond the default boolean flags.