# Tencent Youtu / 腾讯优图 MCP > Tencent Youtu / 腾讯优图 provides professional-grade computer vision, giving your agent instant facial recognition and analysis. Detect faces in images, calculate identity similarity scores, check liveness, and manage secure person libraries—all from one authorized source. ## Overview - **Category:** industry-titans - **Price:** Free - **Tags:** computer-vision, facial-recognition, biometrics, image-analysis, liveness-detection, person-management ## Description Complex image analysis used to verify identities or audit content used to be a nightmare of specialized software logins and manual comparison sheets. Now, your agent handles it all. You can connect this MCP to your preferred AI client and simply talk through visual intelligence workflows. Your agent detects faces in image URLs, pulls out details like age or beauty scores, and compares identities across multiple photos. It even performs static liveness checks for compliance needs. If you're building a Know Your Customer (KYC) system or auditing digital assets, the platform manages person libraries and groups without you ever having to open the Tencent Cloud Console. Using Vinkius, your agent acts as an instant computer vision specialist, giving fast, accurate results directly within your workflow. ## Tools ### analyze_face Analyzes an image and returns detailed metrics about the face, such as estimated age, gender, and beauty score. ### compare_faces Compares two specific faces and generates a numerical similarity score to determine identity matches. ### detect_face Scans an image URL and identifies the precise location of any face present in that picture. ### list_face_groups Retrieves a list of all defined person groups within your registered system library. ### list_persons_in_group Pulls a list of every individual registered and associated with a specific group name. ### live_face_check Verifies if the person in an image is genuinely alive and present, preventing fraud from static media. ### register_person Adds a brand new individual to your secure library by associating them with a photo and group. ### search_face_in_groups Searches through the entire person library, looking for matches based on specific facial features or groups. ## Prompt Examples **Prompt:** ``` Detect faces in this image: [URL] and show me their attributes. ``` **Response:** ``` I've analyzed the image. I detected one face belonging to a male, approximately 28 years old, with a beauty score of 82/100. Should I search for this person in your registered groups? ``` **Prompt:** ``` Compare these two face images: [URL_A] and [URL_B]. ``` **Response:** ``` The facial comparison is complete. The similarity score is 94.5, indicating a very high probability that these are the same person. Would you like me to check if this identity is already registered in your person library? ``` **Prompt:** ``` Register a new person 'Alice' in group 'staff_01' with image [URL]. ``` **Response:** ``` Alice has been successfully registered in group `staff_01` with ID `alice_8821`. Her facial features have been indexed for future recognition. Should I list all members currently in this group? ``` ## Capabilities ### Analyze facial attributes The system processes a face and returns detailed metrics like gender, estimated age, or beauty scores. ### Compare two identities It calculates a precise similarity score between two uploaded images to determine the likelihood that they belong to the same person. ### Identify faces in an image The system scans any image URL and reports exactly where faces are located within it. ### Verify live presence It performs a liveness check on a face to confirm the person is physically present, not using a photo or video loop. ### Manage user groups and records You can list existing groups of people or retrieve specific members registered within those defined groups. ## Use Cases ### Auditing Content for Copyright and Identity A content moderator needs to check thousands of uploaded images for faces belonging to a restricted group. Instead of downloading everything and using desktop software, the agent detects faces in image URLs and uses `search_face_in_groups` to flag any matches within the designated 'restricted' person library. ### Automating Customer Onboarding (KYC) A new user submits a profile photo. The agent first runs the `live_face_check` to confirm identity, then uses `analyze_face` to gather demographic data, and finally calls `register_person` to index them into the 'active_clients' group. ### Investigating Suspects in a Breach A security team needs to know if an unknown face is related to any known bad actors. They use `detect_face` on the suspect image, then run it through `compare_faces` against all records retrieved via `list_persons_in_group` to find potential matches. ### Maintaining Employee Records The HR department needs a current list of staff members in the 'field_sales' group. They ask the agent to use `list_persons_in_group`, instantly getting a vetted, up-to-date roster without touching any backend dashboard. ## Benefits - Eliminate manual image review. Instead of logging into a separate console, your agent uses the `detect_face` tool to find every face in an image and provide immediate attributes. - Speed up compliance checks. Use `live_face_check` whenever you need proof that a person is physically present, making your KYC process much faster than manual video review. - Simplify identity matching. The `compare_faces` tool instantly tells you if two photos are of the same person with a numerical score, eliminating guesswork in asset audits. - Centralize data management. Instead of keeping spreadsheets for who belongs where, use `list_persons_in_group` and `register_person` to manage your entire identity database within your agent's conversation flow. - Streamline search operations. If you need to audit assets, the `search_face_in_groups` tool lets you pinpoint specific people across massive libraries without running complex queries. ## How It Works The bottom line is that it lets you turn complicated biometrics into simple conversational commands. 1. Subscribe to this MCP and enter your necessary Tencent Cloud SecretId and SecretKey. 2. Connect the credential set to your AI client (like Claude or Cursor). 3. Tell your agent what you need. For example, 'Analyze face attributes for this image' or 'Compare these two photos.' The system runs the check and gives you a direct answer. ## Frequently Asked Questions **How does Tencent Youtu / 腾讯优图 handle identity verification?** It compares two faces using the `compare_faces` tool, which generates a specific similarity score. A high score indicates a very strong probability that both photos depict the same person. **Can I use Tencent Youtu / 腾讯优图 to check if someone is real?** Yes, you can run a liveness check using `live_face_check`. This confirms whether the face in the picture or video stream belongs to a living person and isn't just a static image. **What if I need to add new people to my system?** You use the `register_person` tool. This function lets you create a brand-new identity record, linking them to specific photos and assigning them to an official person group. **Does Tencent Youtu / 腾讯优图 help with large datasets?** Absolutely. You can use `search_face_in_groups` to scan through massive person libraries and automatically find matching faces based on criteria or groups you define. **Do I need technical skills to use Tencent Youtu / 腾讯优图?** No. Because it's connected via MCP, you simply ask your agent what you need—for example, 'Analyze face attributes for this image.' The tool handles the complexity.