# EyePop.ai MCP MCP > EyePop.ai brings computer vision to your agents. It lets you analyze images and videos using pre-trained models for object detection, face recognition, and visual classification. You can programmatically manage entire visual pipelines, getting structured metadata like bounding boxes and confidence scores directly through conversation. ## Overview - **Category:** image-video - **Price:** Free - **Tags:** computer-vision, object-detection, face-recognition, visual-intelligence, media-streams, api-integration ## Description Think about manually reviewing video footage or analyzing thousands of photos—it’s slow, tedious work. This MCP lets your agent take over the heavy lifting. Instead of watching every frame, you just point it at the media, and it handles the visual intelligence. You can ask it to list all active visual pipelines for a whole site, track specific objects in real time, or even get detailed metadata about everything it spots. It's like having a dedicated computer vision architect hooked up to your agent via Vinkius. When you need to monitor status or check if the connection is working, it tells you immediately. No more missing critical object detections; your AI acts as the coordinator. ## Tools ### analyze_image Runs object recognition on a single photo, returning bounding boxes and labels for everything it finds. ### analyze_video Processes video files to identify and track objects over time in the stream. ### check_eyepop_status Verifies if your API connection to EyePop is currently active and working correctly. ### create_pop Sets up a new visual processing pipeline for monitoring specific types of media streams. ### get_account Retrieves basic account details to confirm your user setup information. ### get_model Fetches detailed information about a specific visual detection model. ### get_pop Retrieves the current details and status of an existing visual pipeline. ### list_detections Gets a list of all historical object detections that have been recorded by your system. ### list_models Provides a catalog of all the available visual models you can use for detection. ### list_pops Lists every active and inactive visual pipeline set up in your account. ## Prompt Examples **Prompt:** ``` List all active Pops in my EyePop.ai account. ``` **Response:** ``` I've retrieved your Pops. You currently have 5 active visual pipelines, including 'Main Security Feed' and 'Store Counter A'. Would you like the detailed detection metadata for any of them? ``` **Prompt:** ``` Show the detected objects from 'Main Security Feed' for the last hour. ``` **Response:** ``` Visual intelligence orchestrated! For Main Security Feed, I've identified 10 object detections in the last hour, including 5 'Persons' and 2 'Vehicles'. I've retrieved the technical confidence metadata for your review. Need help setting an alert for specific objects? ``` **Prompt:** ``` Check the processing status for Pop ID 'pop_123'. ``` **Response:** ``` Operational monitoring orchestrated! For pop_123, the current status is 'RUNNING' and processing 30 frames per second. Your API connection is healthy. Shall I retrieve the detailed resource usage metadata for this pipeline? ``` ## Capabilities ### Analyze still images Analyzes a single picture and returns labels, detected objects, and their precise locations. ### Process video files Runs temporal object detection on entire videos, providing results over time. ### Manage visual pipelines Creates or lists the specific processing streams (Pops) needed for your monitoring setup. ### Retrieve detailed status Checks the real-time operational health of the EyePop API connection and visual processes. ### Get specific resource data Fetches account information, model details, or specific pipeline metadata when you need to verify settings. ## Use Cases ### Auditing a large site The ops manager needs to know what happened in three weeks of video footage. They ask their agent to `list_pops` first, then tell it to run analysis on the feed using `analyze_video`. The system returns a structured list detailing all detected people and vehicles. ### Debugging an alert A developer gets a suspicious object detection in their app. They use their agent to call `list_detections` for that specific feed, retrieving the confidence scores to determine if the detection is reliable enough to act on. ### Setting up a new monitoring zone The retail analyst needs to track flow in a new area. They ask their agent to use `create_pop` with the necessary parameters, establishing a dedicated visual pipeline for that specific camera feed. ### Checking system health Before running an end-of-month report on all visual data, the developer first runs `check_eyepop_status` to ensure their API key hasn't expired and the connection is solid. ## Benefits - You stop manually reviewing footage. Instead, your agent analyzes the media and gives you structured JSON output with precise bounding box coordinates for every detected item. - Monitoring is instant. Use tools like `check_eyepop_status` to quickly verify API connectivity before starting a major analysis run. - It manages complex visual streams automatically. You can use `create_pop` and `list_pops` to build or check your entire detection setup in one conversational turn. - You never lose metadata. By calling `list_detections`, you get a historical record of everything seen, including confidence scores for every object found. - It’s highly scalable. Use `get_account` and other status tools to monitor the volume of processing without worrying about resource limits. ## How It Works The bottom line is that you just talk to your agent, and it manages the complex communication with EyePop.ai behind the scenes. 1. Subscribe to the MCP and pull your EyePop.ai API Key from your dashboard. 2. Connect this MCP to any compatible agent through Vinkius, giving your AI access to vision tools. 3. Ask your agent to perform a specific task, like analyzing video or listing all pipelines. The agent handles the connection details for you. ## Frequently Asked Questions **How do I analyze video with EyePop.ai using analyze_video?** You call `analyze_video` and provide the file path or stream link to your agent. The agent manages the temporal process, returning detection results for every object across the entire video timeline. **What is the difference between list_models and get_model?** `list_models` gives you a catalog of all available visual models (the types). `get_model` retrieves the specific, detailed settings or parameters for one model once you've chosen it. **Can I check if my API key works before running analyze_image?** Yes. Before any big job, run `check_eyepop_status`. This verifies your connection health and ensures the service is operational for you right now. **How do I list all my active visual pipelines? Use list_pops.** To see everything set up, simply use `list_pops`. This gives you a full inventory of every 'Pop' (pipeline) running in your account. **What does running `check_eyepop_status` confirm about my EyePop connection?** It gives a real-time operational status of your API connectivity and service health. This is the best way to verify if your account can handle complex visual tasks before you run them. **How do I set up a new, dedicated visual pipeline using `create_pop`?** You call `create_pop` and provide the necessary configuration details. This initializes a specific, named 'Pop' (Pipeline) within your account for ongoing, structured analysis. **After an image or video run, how do I get the detailed object bounding box coordinates using `list_detections`?** You pass the relevant Pop ID to `list_detections`. The tool returns a structured list containing all detected objects and their precise boundary coordinates (bounding boxes) for data use. **What kind of usage quotas can I view by calling `get_account`?** This function fetches your core profile information alongside critical billing and usage quota details. It helps you monitor service scaling and track remaining credits or limits. **How do I find my EyePop.ai API Key?** Log in to your account, click on your profile, and copy your unique **API Key** from the developer section. **Can I check object detection results via AI?** Yes! The `list_detections` tool allows your agent to retrieve metadata including object classes and confidence scores for any media stream. **How do I list my active Pops?** Use the `list_pops` tool to retrieve your complete directory along with the unique identifiers for all managed visual pipelines.