Skip to main content
Animus provides content moderation capabilities that help you identify and filter potentially harmful content. Unlike some other APIs, our moderation is integrated directly into our chat completions endpoint using the compliance parameter.

​
Using Content Moderation

To enable content moderation in your API requests, simply set the compliance parameter to true in your chat completion requests. When enabled, the API will analyze the content and return details about any detected violations. 
// Example using fetch API
const response = await fetch('https://api.animusai.co/v2/chat/completions', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': `Bearer ${process.env.ANIMUS_API_KEY}`
  },
  body: JSON.stringify({
    model: "vivian-llama3.1-70b-1.0-fp8",
    messages: [
      { role: "user", content: "Text to check for compliance" }
    ],
    compliance: true  // Enable content moderation
  })
});

const data = await response.json();

// Check if there are any compliance violations
if (data.compliance_violations && data.compliance_violations.length > 0) {
  console.log("Content violations detected:", data.compliance_violations);
} else {
  console.log("No content violations detected");
}

​
Response Format

When compliance checking is enabled, the API response will include a compliance_violations field that contains an array of any detected content violations. Here’s an example response with detected violations: 
{
  "id": "chat-abcd1234",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "vivian-llama3.1-70b-1.0-fp8",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Response content here..."
      }
    }
  ],
  "usage": {
    "prompt_tokens": 40,
    "completion_tokens": 60,
    "total_tokens": 100
  },
  "compliance_violations": ["drug_use"]
}

​
Content Violation Categories

Our compliance system can detect and flag the following categories of potentially harmful content:
CategoryDescription
pedophiliaContent related to sexual content involving minors
beastialityContent involving sexual acts with animals
murderContent that promotes or glorifies murder
rapeContent related to sexual assault
incestContent involving sexual relations between family members
goreExplicit and graphic violent content
prostitutionContent promoting or soliciting prostitution
drug_useContent promoting or describing drug use

​
Advanced Implementation

​
Handling Different Violation Types

async function handleContentModeration(userContent) {
  const response = await fetch('https://api.animusai.co/v2/chat/completions', {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'Authorization': `Bearer ${process.env.ANIMUS_API_KEY}`
    },
    body: JSON.stringify({
      model: "vivian-llama3.1-70b-1.0-fp8",
      messages: [
        { role: "user", content: userContent }
      ],
      compliance: true
    })
  });
  
  const data = await response.json();
  
  if (data.compliance_violations && data.compliance_violations.length > 0) {
    // Handle different types of violations
    const violations = data.compliance_violations;
    
    if (violations.includes('drug_use')) {
      console.log("Drug-related content detected");
      // Implement specific handling for drug content
    }
    
    if (violations.includes('gore') || violations.includes('murder')) {
      console.log("Violent content detected");
      // Implement specific handling for violent content
    }
    
    // Log for review
    logViolation(userContent, violations);
    
    return {
      allowed: false,
      violations: violations,
      message: "Content violates our community guidelines"
    };
  }
  
  return {
    allowed: true,
    content: data.choices[0].message.content
  };
}

function logViolation(content, violations) {
  // Log violation for review and analysis
  console.log(`Violation logged: ${violations.join(', ')} - Content: ${content.substring(0, 100)}...`);
}

​
Batch Content Moderation

For applications that need to check multiple pieces of content: 
async function batchModerationCheck(contentArray) {
  const results = await Promise.allSettled(
    contentArray.map(content => 
      handleContentModeration(content)
    )
  );
  
  return results.map((result, index) => ({
    index,
    content: contentArray[index],
    result: result.status === 'fulfilled' ? result.value : { error: result.reason }
  }));
}

// Usage
const contentToCheck = [
  "This is normal content",
  "This might contain violations",
  "Another piece of content to check"
];

const batchResults = await batchModerationCheck(contentToCheck);
batchResults.forEach(({ index, content, result }) => {
  if (result.allowed) {
    console.log(`Content ${index}: Approved`);
  } else {
    console.log(`Content ${index}: Rejected - ${result.violations?.join(', ')}`);
  }
});

​
Best Practices

For effective content moderation in your applications:
  1. Always enable compliance: Set compliance: true for all user-generated content
  2. Implement appropriate responses: Create user-friendly notifications when content is flagged
  3. Pair with frontend filters: Implement basic filtering on the client side to reduce API calls for obvious violations
  4. Handle violations gracefully: Provide constructive feedback to users when their content is flagged
  5. Review edge cases: Periodically review flagged content to understand common violations in your application
  6. Log violations: Keep records for analysis and improvement of your moderation system
  7. Implement appeals process: Allow users to appeal moderation decisions when appropriate

​
Error Handling

Implement robust error handling for moderation requests: 
async function safeContentModeration(content) {
  try {
    const response = await fetch('https://api.animusai.co/v2/chat/completions', {
      method: 'POST',
      headers: {
        'Content-Type': 'application/json',
        'Authorization': `Bearer ${process.env.ANIMUS_API_KEY}`
      },
      body: JSON.stringify({
        model: "vivian-llama3.1-70b-1.0-fp8",
        messages: [{ role: "user", content: content }],
        compliance: true
      })
    });

    if (!response.ok) {
      throw new Error(`HTTP error! status: ${response.status}`);
    }

    const data = await response.json();
    return {
      success: true,
      violations: data.compliance_violations || [],
      content: data.choices[0].message.content
    };

  } catch (error) {
    console.error('Moderation check failed:', error);
    return {
      success: false,
      error: error.message,
      // Fail safe - assume content needs review
      violations: ['moderation_error']
    };
  }
}

​
Integration Examples

​
Web Application Integration

// Example: Chat application with real-time moderation
class ChatModerator {
  constructor(apiKey) {
    this.apiKey = apiKey;
  }

  async moderateMessage(message) {
    const result = await this.checkCompliance(message);
    
    if (!result.allowed) {
      return {
        blocked: true,
        reason: this.getViolationMessage(result.violations),
        violations: result.violations
      };
    }

    return {
      blocked: false,
      content: result.content
    };
  }

  getViolationMessage(violations) {
    const messages = {
      'drug_use': 'Messages about drug use are not allowed.',
      'gore': 'Graphic violent content is not permitted.',
      'murder': 'Content promoting violence is prohibited.',
      // Add more specific messages
    };

    const specificViolations = violations
      .map(v => messages[v])
      .filter(Boolean);

    return specificViolations.length > 0 
      ? specificViolations.join(' ')
      : 'Your message violates our community guidelines.';
  }
}

// Usage in chat application
const moderator = new ChatModerator(process.env.ANIMUS_API_KEY);

async function handleUserMessage(userMessage) {
  const moderationResult = await moderator.moderateMessage(userMessage);
  
  if (moderationResult.blocked) {
    showUserWarning(moderationResult.reason);
    return;
  }

  // Process the approved message
  displayMessage(moderationResult.content);
}

​
Next Steps

Text Generation

Learn how to generate text with built-in moderation

Vision

Understand vision capabilities and content analysis

Webhooks

Set up webhooks for automated moderation workflows

API Reference

Complete API documentation and reference
⌘I