Manual Moderation
Human expert review for nuanced content moderation decisions
Manual moderation provides human expert review for content that requires cultural context, nuanced judgment, or complex policy decisions that AI might miss. Perfect for edge cases and high-stakes content decisions.
What is Manual Moderation?
Manual moderation is a human-in-the-loop content review system where trained moderators analyze content submissions and make informed decisions based on your platform's policies. Unlike automated moderation, manual review can understand context, cultural nuances, and complex scenarios that require human judgment.
When you submit content for manual moderation, it's added to a review queue where human moderators can examine the full context, apply policy knowledge, and make nuanced decisions. You'll receive webhook notifications when reviews are complete.
How Manual Moderation Works
Review Workflow
Content Submission
Your application sends content to the manual moderation API endpoint using the same schema structure as automated moderation.
Queue Assignment
Content is added to the manual moderation queue where trained moderators can access it for review.
Human Review
Moderators examine the content against your enabled categories, consider context and policies, and make informed harmful/not harmful decisions. Reviews are typically completed within 48 hours.
Webhook Notification
When the review is complete, you receive a webhook notification with the moderation decision and details.
When to Use Manual Moderation
✓Best Use Cases
- • Borderline content requiring context analysis
- • Cultural or linguistic nuance evaluation
- • High-stakes decisions with policy implications
- • Appeals and dispute resolution
- • Content that combines multiple categories
- • Platform-specific policy enforcement
⚠Consider Carefully
- • Large volumes of straightforward content
- • Time-sensitive moderation decisions
- • Clear-cut violations easily caught by AI
- • Budget-constrained scenarios
- • Simple binary policy rules
Configuration
Manual moderation requires a subscription and webhook configuration to receive notifications when moderation decisions are complete. You can manage these settings through the Console.
Configuration Steps
- 1Subscribe to Manual Moderation
Enable manual moderation for your project through the Console billing section
- 2Configure Webhook URL
Set your endpoint URL to receive moderation decision notifications
- 3Test Webhook
Use the test webhook feature to verify your endpoint receives notifications correctly
- 4Start Submitting Content
Use the manual moderation API endpoint to submit content for human review
Webhook Notifications
When manual moderation is complete, you'll receive a webhook notification with the moderation decision. This allows your application to react to moderation results asynchronously. The webhook payload includes the final decision, moderator notes, and detailed component-level results.
Webhooks are essential for manual moderation since review times can vary from minutes to hours depending on content complexity and queue volume. Your application should be prepared to handle these asynchronous notifications.
Webhook Security
Webhooks include signature headers for verification and should be configured with secrets for payload validation. Always verify webhook authenticity and use HTTPS endpoints for secure transmission.
Combining with Automated Moderation
Many platforms use a hybrid approach, combining automated and manual moderation for optimal results. This provides both speed and accuracy while managing costs effectively.
🤖AI First
Use automated moderation for initial screening to catch obvious violations quickly and cost-effectively.
Handle 80-90% of content automatically
⚖️Human Review
Route borderline cases, appeals, and nuanced content to manual moderation for expert review.
Focus on complex 10-20% of cases
✨Best Results
Achieve high accuracy with reasonable costs by leveraging the strengths of both approaches.
Optimal balance of speed, cost, and quality
Best Practices
Submission Guidelines
- • Provide sufficient context in your content
- • Use descriptive schema component names
- • Include relevant metadata when possible
- • Batch similar content types together
Webhook Handling
- • Implement idempotent webhook processing
- • Store submission IDs to track status
- • Handle webhook delivery delays gracefully
- • Verify webhook signatures for security
Cost Optimization
- • Use automated moderation for initial filtering
- • Reserve manual review for complex cases
- • Monitor moderation patterns and adjust
- • Consider batch processing for efficiency
Quality Assurance
- • Review moderation decisions periodically
- • Provide feedback on decision quality
- • Update policies based on edge cases
- • Monitor response times and patterns
Subscription & Pricing
Manual moderation requires a subscription and is priced per reviewed submission. This reflects the human expertise and time involved in thorough content analysis. See our pricing plans for detailed cost information.
Pricing Considerations
- • Per-submission pricing based on content complexity
- • Volume discounts available for large-scale usage
- • No charges for submissions that remain in queue
- • Transparent billing with detailed usage reports
Ready to Get Started?
Set up manual moderation for your project and start leveraging human expertise for nuanced content decisions.
Related Documentation
- • Automated Moderation - AI-powered content analysis
- • Schemas & Components - Structure your content for moderation
- • Console Walkthrough - Configure webhooks and subscription
- • Pricing Plans - Manual moderation subscription options