Skip to main content

​
Content Rating (Alpha)

Let’s face it: not every image is meant for every audience. Whether you’re building a community-driven app, a professional tool, or just want to keep things organized, knowing what’s actually in your generated images is crucial. Enter Content Rating. 🛡️ Instead of manually squinting at thousands of pixels, you can let our AI sit in the curator’s chair. It analyzes every image the moment it’s born and gives you a definitive rating based on your rules.
Alpha Stage: Content Rating is currently in early alpha. We use sophisticated vision models to analyze images, but they can occasionally make mistakes. Always have a human in the loop for critical moderation.

​
What is it?

Content Rating is an optional post-processing step in our image generation pipeline. When enabled, we take the final image and pass it through a dedicated Vision AI model specifically trained for content moderation. It doesn’t just guess; it understands context. It can tell the difference between a “statue in a museum” (usually SFW) and something that definitely isn’t.

​
Why use it?

​
1. Part of the Pipeline

Just like Magic Prompt, Content Rating is integrated directly into the Mynth workflow. Since the analysis happens during the generation lifecycle, it is included in your standard service fee—allowing you to maintain high standards for your content without worrying about additional per-image costs.

​
2. Custom Moderation Logic

Every project has different standards. While we provide a sensible “SFW/NSFW” default, you can define your own rating levels. Want to flag images as “High Fantasy,” “Photo-realistic,” or “Cringe”? You can do that.

​
3. Automated Workflows

By getting the rating directly in the API response, you can automate your next steps. Auto-blur NSFW content, move specific styles to different folders, or trigger alerts—all without lifting a finger.

​
Behavior & Control

Content Rating is disabled by default to keep things lighting-fast. To enable it, you just need to add a small flag to your API request.

​
Basic Enablement

To use the default sfw / nsfw logic: 
{
  "content_rating": {
    "enabled": true
  }
}

​
Custom Levels

If the default binary isn’t enough, you can define up to 7 custom levels. Each level needs a value (what you’ll get back) and a description (how the AI should decide). 
{
  "content_rating": {
    "enabled": true,
    "levels": [
      { "value": "safe", "description": "Generic content suitable for all ages." },
      { "value": "suggestive", "description": "Borderline content, mild themes." },
      { "value": "explicit", "description": "Adult content, nudity, or violence." }
    ]
  }
}
Pro Tip: Keep your descriptions short and clear. Instead of “Anything that looks like a cat but is actually a dog,” use “Canine-focused imagery.”

​
The Result

The rating is returned alongside each image in the task result. It’s clean, predictable, and ready for your code. 
{
  "images": [
    {
      "url": "https://cdn.mynth.io/...",
      "content_rating": {
        "mode": "default",
        "level": "sfw"
      }
    }
  ]
}

Roadmap: Intelligence Beyond Rating 🧠

Content Rating is just the beginning. Our mission is to offload the tedious manual work of asset management from you and your developers.What’s coming next:
  • Automated Tagging: A sophisticated categorization system to index your images by style, subject, and mood automatically.
  • AI Titles & Descriptions: High-quality metadata generation to make your galleries searchable and accessible without any manual input.
We’re listening! Have a specific use case or a feature idea? Join our Discord and let us know. We want to build the tools that make your life easier.
So go ahead, let the AI do the heavy lifting. Your eyes (and your users) will thank you.