- EMOTION
- GESTURE
- POSTURE
- HUMAN
Short answer
How can I use my AI to track user behavior?
Your agent uses the Face++ MCP to extract facial expressions, hand gestures, and body poses from images. This turns visual data into text-based descriptions of how people are acting. You'll see the specific outcomes below.
The results you get.
Behavioral data outputs.
The AI processes your images into these distinct categories of information.
EMOTION
Facial expression data
The AI identifies specific facial movements. You get a clear sense of the user's emotional state.
GESTURE
Hand and arm tracking
Your agent detects specific hand signals. This helps you understand non-verbal communication.
POSTURE
Body language summaries
The AI maps the physical stance of the subject. You receive a description of their overall posture.
HUMAN
Manual review needed
The AI provides the raw behavioral data. You must interpret these findings to make final decisions.
The workflow
What your AI does when the image arrives.
The AI follows a logical path to turn pixels into human insights.
Locate subjects
The AI scans the image to find all human faces present.
detect_faceIdentify gestures
The agent looks for specific hand and arm movements.
gesture_detectMap posture
It tracks the physical structure and position of the body.
skeleton_detectSummarize behavior
The AI combines all detected cues into a single behavioral description.
detect_body
Try it
Copy these to start.
Use these prompts to begin your analysis.
Starting points
Swap the bracketed text for your actual image files or specific video frames.
Face++ / Megvii Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_9XeNBogqnvbfe37ielQPllFDO6AZEm0YUpuNS984/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Face++ / Megvii capabilities are ready to use.
{
"mcpServers": {
"face-megvii-mcp": {
"url": "https://edge.vinkius.com/vk_preview_9XeNBogqnvbfe37ielQPllFDO6AZEm0YUpuNS984/mcp"
}
}
}Analyze this image and tell me what emotions the people are showing.
Look at the person in this photo and describe their hand gestures.
Check the body language in this video frame to see if the user looks frustrated.
Map the skeleton and body pose of the subject in this file to understand their stance.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Face++ once, then ask.
Link the MCP to your client and you're ready. Your login stays encrypted, and you can start asking questions immediately.
Connect Face++ to your AIFAQ
How this task behaves.
- 01
Can the AI identify people by name?
No. The tools detect facial features and movements, but they do not perform identity recognition.
- 02
Can the AI change the image?
No. The AI only reads the image data to interpret behavior. It cannot edit or modify the files.
- 03
Does the AI track movement in real-time?
The AI processes the specific images or video frames you provide to it.
- 04
Can the AI tell if someone is lying?
No. The AI reports on detected gestures and expressions. It cannot perform psychological assessments.
- 05
Will the AI save my images?
No. The AI only uses the images you pass to it through your client to perform the analysis.
More questions about Face++? The Connector page answers them. See everything the Face++ Connector can do
Connect Face++ / Megvii to Claude, Cursor, ChatGPT & more
