# Hive AI MCP for AI Agents AI Agent Connect > Hive AI handles content moderation and AI detection for your AI agents. It filters hate speech, violence, and NSFW content from text, images, and video in real-time. Use it to identify AI-generated content or manage large-scale moderation tasks asynchronously to keep your community safe. ## Overview - **Category:** ai-frontier - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_67vdfEmExljNvTYJ0XEqiEcUEkFjgnhergEQzSF5/ai-agent-connect - **Tags:** content-moderation, ai-detection, nsfw-filtering, safety-compliance, real-time-analysis ## Description Hive AI connects your safety protocols directly to your AI agent. Instead of manually reviewing every piece of user-generated content, you can have your agent act as a dedicated safety lead. It handles the heavy lifting of scanning text for hate speech or checking images for NSFW material before they ever go live. If you're dealing with large files like videos, it manages those in the background so you don't have to wait around for a response. It also helps you figure out which content was actually made by a human and which came from a generative model. It's a big part of why many teams use the Vinkius catalog to build out their safety stacks. You get a clear picture of your platform's safety health with detailed scores and categories that tell you exactly why something was flagged. This means you can move faster without worrying about your community guidelines getting ignored. It gives your agent the specific knowledge it needs to make smart calls on what stays and what goes. ## Tools ### get_async_task_result Fetch the final moderation results for a completed background task. Use this once your video or audio file has finished processing. ### detect_ai_generated_image Check if an image was made by a tool like Midjourney or DALL-E. It returns a probability score for AI generation. ### detect_ai_generated_text Determine if a block of text was written by an AI model. This helps you spot automated spam or fake reviews. ### get_project_details View the specific configuration and settings for your Hive AI project. It's useful for double-checking your current project scope. ### get_async_task_status Check if a background moderation task is still running or finished. Provide the unique task ID to see current progress. ### list_available_models See which Hive AI models are active for your project. This helps you pick the right model for different content types. ### moderate_audio_async Start a moderation task for an audio file and get a task ID back. This handles long files without blocking your agent. ### moderate_image Perform real-time moderation on a public image URL. It checks for safety violations instantly. ### moderate_text Verify user-generated text for safety before it gets published. It checks for things like hate speech or violence. ### moderate_video_async Submit a video for deep moderation and receive a task ID. Use this for content that requires more than a quick scan. ## Prompt Examples **Prompt:** ``` Check this comment: 'I hate everyone and want to hurt people.' Is it safe? ``` **Response:** ``` I've analyzed that comment. **It is not safe.** Here are the results: - **Hate Speech:** 94% confidence - **Violence:** 88% confidence This content violates safety standards and should be blocked immediately. Would you like me to flag the user profile for further review? ``` **Prompt:** ``` Is this image a deepfake? https://example.com/photo.jpg ``` **Response:** ``` Scanning the image now... **Result:** High probability of AI generation. - **AI Generated:** 98.2% - **Human Created:** 1.8% This image shows strong signs of being created by a tool like Midjourney. Do you want me to check other images from this user? ``` **Prompt:** ``` I have a video at https://example.com/clip.mp4. Can you check it for safety? ``` **Response:** ``` I've started the video moderation task. - **Task ID:** `vid_7721_x` - **Status:** Queued Deep analysis for videos can take a few minutes. I'll monitor the progress and let you know as soon as the safety report is ready. ``` ## Capabilities ### Filter hate speech and NSFW content Your agent can scan text and images to block violations instantly. ### Detect synthetic media Identify if a piece of content was made by a human or a generative AI model. ### Process large video files Submit videos for deep moderation in the background without blocking your agent. ### Track background tasks Monitor the status of long-running moderation jobs using unique IDs. ### View model configurations See which Hive AI models are active and how your projects are set up. ### Get safety scores Access detailed scores to prove your platform meets safety and compliance rules. ## Use Cases ### Blocking toxic comments in real-time A community manager wants to block a toxic comment. They ask the agent to moderate the text, and it flags the hate speech immediately. ### Verifying user profile photos A platform owner wants to know if a user's profile pic is a deepfake. They use `detect_ai_generated_image` to check the URL. ### Scanning long-form video content A developer needs to scan a 5-minute video for NSFW content. They use `moderate_video_async` to get a background task started. ### Generating weekly safety reports A trust team needs a weekly report on safety. They use `get_project_details` and `get_async_task_result` to gather data on flagged items. ## Benefits - Stop manual reviews by using `moderate_text` to filter hate speech and violence automatically before it hits your feed. - Protect your brand by using `detect_ai_generated_image` to find and flag synthetic media that could mislead your users. - Handle heavy media like video with `moderate_video_async` so your agent doesn't time out while waiting for results. - Get clear safety scores to prove your platform meets compliance standards for your stakeholders and advertisers. - Save time on multi-modal content by using `moderate_audio_async` to scan for sensitive audio in your community. - Keep your project organized by using `list_available_models` to ensure your agent uses the right tools for the job. ## How It Works The bottom line is your AI agent takes over the manual labor of content policing. 1. Subscribe to the Hive AI MCP and add your API keys. 2. Tell your agent which content to check or which project to use. 3. Get instant safety scores or background task IDs for media analysis. ## Frequently Asked Questions **Can Hive AI moderate my community's chat?** Yes, it can. Your AI agent can scan every message for hate speech or violence and flag violations automatically before they reach other users. **How does Hive AI detect if someone used Midjourney?** It uses specialized detection models to analyze images and return a probability score of whether the content was generated by AI. **Can I use Hive AI to moderate videos?** Yes, it supports asynchronous video moderation. This allows your agent to submit a video for deep analysis and check back later for the results. **Does Hive AI provide safety scores for my reports?** It provides detailed safety scores and categories, making it easy to generate compliance reports for your stakeholders or advertisers. **Can I see which Hive AI models are available?** Yes, your agent can list all the Hive AI models currently available for your specific project so you can choose the best fit for your content. **How do I handle long moderation tasks with Hive AI?** For long files like audio or video, the Connector provides a task ID. Your agent can use this ID to check the status and retrieve results once finished. **How do I find my Hive AI API Keys?** Log in to your Hive AI dashboard, select a project (e.g., a Visual or Text project), and navigate to the **Settings** or **API** section. Note that each project type requires its own unique API key. **What is the difference between Synchronous and Asynchronous moderation?** Synchronous moderation (`moderate_text`, `moderate_image`) returns results instantly and is best for real-time interactions. Asynchronous moderation (`moderate_video_async`) is used for larger files and requires polling the task status or using a callback. **Can this integration detect deepfakes or AI-generated voices?** Yes! Hive AI provides specialized models for detecting AI-generated content across text, images, video, and audio. Use the `detect_ai_generated_text` and `detect_ai_generated_image` tools for these checks. **Is the integration secure for sensitive content?** Absolutely. The integration uses official Hive AI Token authentication over HTTPS. Your credentials and analyzed content are encrypted and stored securely within the Vinkius Cloud infrastructure.