# Tuputech Moderation MCP for AI Agents AI Agent Connect > Tuputech Moderation MCP brings advanced content safety to your AI workflows. It automatically scans text, images, and audio for things like deepfakes, porn, and spam using Tuputech's machine-learning database. Use it to build trust and safety layers into your apps without manual review. ## Overview - **Category:** image-video - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_gb0XPkdhb9Za66CEymV09DgwDJyFdD3cFgotR1xY/ai-agent-connect - **Tags:** content-moderation, deepfake-detection, anti-spam, visual-recognition, trust-and-safety ## Description Imagine you're running a community where moderation is a constant headache. You're constantly worried about NSFW content, deepfakes, or spam bots flooding your feeds. This Connector lets your AI handle that heavy lifting for you. Instead of your agent trying to guess if an image is safe based on its training data, it can query a dedicated evaluation database that's specifically built for these risks. It handles everything from identifying synthetic media to flagging toxic language in real-time. You can set up automated filters for video streams or audio fragments so you don't have to worry about what's being uploaded. It's a way to ground your AI's judgment in actual safety standards. Finding this in the Vinkius catalog makes it easy to plug into your existing setup and start filtering content the moment it hits your platform. It moves the responsibility of safety from a best guess to a data-driven check. ## Tools ### scan_image_ad Looks for hidden advertisements inside images to keep your feed clean. It helps identify marketing content that bypasses text filters. ### scan_image_porn Scans images for NSFW or pornographic content to keep your app safe. It provides a high-confidence check for explicit material. ### scan_image_synthetic Detects if an image is a deepfake or AI-generated. Use this to spot synthetic media that tries to impersonate real people. ### scan_image_terrorism Checks images for gore or terrorism-related content. This helps maintain strict safety standards for public platforms. ### scan_audio_stream Starts an asynchronous check on audio files to find safety violations. This is useful for monitoring voice clips or live streams. ### scan_text_abuse Identifies toxic language and verbal abuse in text blocks. It helps keep community chats friendly and free of harassment. ### submit_feedback Sends manual feedback to improve the accuracy of the moderation results. Use this to refine the AI's understanding of your rules. ### scan_text_politics Checks if text violates specific political guidelines or rules. This is useful for maintaining neutral or moderated spaces. ### scan_text_spam Filters out spam and advertisements from user comments. It keeps your engagement metrics clean from bot activity. ### scan_video_stream Initiates an asynchronous check on video files for prohibited content. This allows you to scan large files without blocking the user. ## Prompt Examples **Prompt:** ``` Check if this profile photo contains NSFW or Gore imagery: https://example.com/photo.jpg ``` **Response:** ``` I've scanned the image using the safety database. Here are the results: | Category | Status | Confidence | | :--- | :--- | :--- | | **Pornography** | Passed | 99.4% | | **Terrorism/Gore** | Passed | 98.8% | The image is safe to display on your platform. ``` **Prompt:** ``` Scan this text and see if it is spam or political botting: 'Buy followers here...' ``` **Response:** ``` I've checked the text against the spam and political filters. **Result:** Rejected **Reason:** The content was flagged as **Spam/Fraud**. It contains common patterns used in bot-driven marketing. ``` **Prompt:** ``` Is this video safe for a general audience? https://example.com/video123 ``` **Response:** ``` I've initiated an asynchronous scan on the video stream. * **Status:** Processing * **Action:** The agent will notify you once the safety check is complete. I'll check back in a moment to see if it's cleared for the general feed. ``` ## Capabilities ### Detect deepfakes Identify if an image was made by AI or a human. ### Block NSFW content Scan images for pornographic or inappropriate material. ### Filter spam Catch advertisements and spam in text blocks automatically. ### Monitor audio streams Analyze audio fragments for safety violations. ### Flag toxic language Detect abuse or hate speech in user comments. ### Identify synthetic media Spot AI-generated images before they go live. ### Scan video content Run asynchronous checks on video files for safety. ## Use Cases ### Community Safety A Discord bot uses scan_text_spam to block bot-driven links in a chat, preventing users from being led to phishing sites. ### App Moderation A social app uses scan_image_porn to instantly reject inappropriate profile photos during the signup flow. ### Content Review A video platform uses scan_video_stream to flag gore or NSFW clips for human review, keeping the feed safe for all ages. ### Brand Protection A marketing tool uses scan_image_ad to ensure user posts don't contain unauthorized ads or deceptive marketing. ## Benefits - Stop deepfakes from spreading by using scan_image_synthetic to catch AI-generated media before it goes live. - Keep your community clean with scan_text_abuse which flags toxic language before it can be seen by others. - Save on storage costs by using scan_image_porn to block NSFW avatars before they ever hit your CDN. - Handle high-volume feeds with scan_video_stream for asynchronous safety checks that don't slow down your app. - Protect your brand from spam by using scan_image_ad to find hidden marketing in user-submitted images. ## How It Works The bottom line is you get a dedicated safety layer that handles complex content moderation automatically. 1. Get your Access Key (SecretId) and generate your RSA Private Key from the Tuputech Developer Console. 2. Input these credentials into the Connector configuration. 3. Your agent will now handle the RSA hashing internally and query the database directly for safety scores. ## Frequently Asked Questions **Can Tuputech Moderation MCP detect deepfakes?** Yes, it specifically includes tools to identify if an image was generated by AI or a human, which is essential for stopping synthetic media from spreading. **How does it handle video and audio?** It provides asynchronous scanning for both video and audio streams. This allows your agent to check large files without blocking your app's performance. **Can I use it for a Discord bot?** Absolutely. It's perfect for community bots that need to automatically filter spam, toxic language, and prohibited links in real-time. **Does it block specific types of images like porn?** Yes, it has dedicated scanning for NSFW and pornographic content, helping you keep your platform safe and compliant with safety standards. **How accurate is the spam detection?** It uses a dedicated machine-learning database designed for high-accuracy detection of advertisements and spam in both text and images. **Can I submit feedback to improve results?** Yes, the Connector includes a feedback tool that lets you submit manual corrections to help refine the moderation accuracy over time. **How is the RSA authentication performed?** No manual prompt instructions are needed. You simply set your RSA Private Key safely into your Connector profile in Vurb. When your AI triggers an endpoint like `scan_image_porn`, the Vurb NodeJS shell dynamically computes the RSA `signature` hashing your exact Nonce+Timestamp completely isolated from your prompt traces.