🤖

Automated Moderation API

Submit content for instant AI-powered moderation

The automated moderation endpoint provides instant AI-powered content analysis. Submit content with a schema and receive immediate harmful/safe classifications with detailed category breakdowns.

Submit Content

POSThttps://api.outharm.com/moderation/automated

Submits content for immediate AI moderation and returns detailed results.

📋 Requirements

  • • Authentication: Bearer token required
  • • Content-Type: application/json
  • • Subscription: Automated moderation access or free trial credits

Request Body

Request Structure

{
  "schema_id": "550e8400-e29b-41d4-a716-446655440000",
  "content": {
    "field_name": ["value1", "value2"],
    "another_field": ["value3"]
  }
}

Field Definitions

schema_idREQUIREDUUID

The ID of the schema that defines the content structure. Must belong to your project and define the expected content fields.

contentREQUIREDObject

Content to moderate, organized by field names. Field names must match schema component names. All values are arrays for consistency.

Content Value Types
Text Values:
"title": ["Blog post about AI safety", "Another title"]
Image URLs:
"images": ["https://example.com/image.jpg", "https://example.com/photo.png"]
Base64 Images:
"avatar": ["data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD..."]

⚠️ Validation Rules

Content Structure:

  • • Content object cannot be empty
  • • Maximum 50 fields per request
  • • Field names must match schema component names
  • • Field names cannot be empty or exceed 100 characters

Content Arrays:

  • • Arrays cannot be empty (must have at least one value)
  • • Maximum 100 values per field array
  • • Individual values cannot be empty or whitespace-only

URLs & Images:

  • • HTTP/HTTPS URLs must be under 2048 characters
  • • Base64 data URLs must include `;base64,` format

Request Examples

Text Content Moderation

POST https://api.outharm.com/moderation/automated
Authorization: Bearer your_api_token_here
Content-Type: application/json

{
  "schema_id": "550e8400-e29b-41d4-a716-446655440000",
  "content": {
    "title": ["Check this post title for harmful content"],
    "description": [
      "This is a user-generated post description.",
      "Another description to check simultaneously."
    ],
    "tags": ["technology", "ai", "moderation"]
  }
}

Mixed Content (Text + Images)

{
  "schema_id": "550e8400-e29b-41d4-a716-446655440000",
  "content": {
    "title": ["Product listing with suspicious content"],
    "description": ["Check this product description"],
    "images": [
      "https://example.com/product-image.jpg",
      "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD..."
    ],
    "category": ["electronics"]
  }
}

Response

Response Structure

{
  "submission_id": "123e4567-e89b-12d3-a456-426614174000",
  "schema_id": "550e8400-e29b-41d4-a716-446655440000",
  "is_harmful": true,
  "results": {
    "field_name": {
      "is_harmful": true,
      "detailed": [
        {
          "is_harmful": true,
          "categories": ["violence", "hate"]
        },
        {
          "is_harmful": false
        }
      ]
    }
  }
}

Response Fields

submission_idUUID

Unique identifier for this moderation request. Use this to track or reference the submission.

schema_idUUID

The schema ID that was used for this moderation (same as request).

is_harmfulBoolean

Overall harmful status across all submitted content. true if any content was detected as harmful.

resultsObject

Detailed moderation results organized by field name (component). Each field contains aggregated and detailed results.

Component Result Structure
is_harmfulBoolean

Whether this field (component) contains any harmful content.

detailedArray

Individual results for each value in the field array. Array order matches request order.

Item Result Structure:

is_harmfulBoolean

Individual item harmful status

categoriesString[] | null

Harmful categories detected (only present if is_harmful is true)

Response Examples

Safe Content Response

{
  "submission_id": "123e4567-e89b-12d3-a456-426614174000",
  "schema_id": "550e8400-e29b-41d4-a716-446655440000",
  "is_harmful": false,
  "results": {
    "title": {
      "is_harmful": false,
      "detailed": [
        {
          "is_harmful": false
        }
      ]
    },
    "description": {
      "is_harmful": false,
      "detailed": [
        {
          "is_harmful": false
        },
        {
          "is_harmful": false
        }
      ]
    },
    "tags": {
      "is_harmful": false,
      "detailed": [
        {
          "is_harmful": false
        }
      ]
    }
  }
}

Harmful Content Response

{
  "submission_id": "123e4567-e89b-12d3-a456-426614174000",
  "schema_id": "550e8400-e29b-41d4-a716-446655440000",
  "is_harmful": true,
  "results": {
    "title": {
      "is_harmful": true,
      "detailed": [
        {
          "is_harmful": true,
          "categories": ["violence", "hate"]
        }
      ]
    },
    "description": {
      "is_harmful": false,
      "detailed": [
        {
          "is_harmful": false
        },
        {
          "is_harmful": false
        }
      ]
    },
    "images": {
      "is_harmful": true,
      "detailed": [
        {
          "is_harmful": false
        },
        {
          "is_harmful": true,
          "categories": ["sexual", "violence"]
        }
      ]
    }
  }
}

Error Responses

Validation Errors (400 Bad Request)

When request validation fails:

{
  "name": "validation-error",
  "message": "Content cannot be empty",
  "status_code": "400"
}

// Other validation examples:
{
  "name": "validation-error", 
  "message": "Field 'title' cannot be an empty array",
  "status_code": "400"
}

{
  "name": "validation-error",
  "message": "Too many fields (max 50)",
  "status_code": "400"
}

Schema Errors (404 Not Found)

When schema is not found or not accessible:

{
  "name": "schema-not-found",
  "message": "Schema not found or not accessible",
  "status_code": "404"
}

Subscription Errors (403 Forbidden)

When automated moderation access is required:

{
  "name": "automated-moderation-subscription-required",
  "message": "Automated moderation subscription required",
  "status_code": "403"
}

Harmful Categories

When content is detected as harmful, the response includes category names that identify the type of harmful content. See Categories Documentation for the complete list of available categories and their descriptions.

📌 Category Examples

Common category names returned in responses:

violencehatesexualharassmentself-harmillegalspamtoxicprofanity

Ready to Get Started?

Start moderating content with the automated API. Create schemas, configure categories, and begin protecting your platform.

Create Schemas

Define your content structure

→

Configure Categories

Choose what to detect as harmful

→

Related Documentation