• DETECTION
  • COORDINATES
  • DATA
  • HUMAN

How to identify and locate objects in images with your AI

Connect NVIDIA Vision to your AI client to turn photos into structured data. Once connected, your agent can see what's in a frame and where it sits.

Ask AI about this page

Short answer

How can my AI find things in a photo?

Your AI uses NVIDIA Vision to scan images, identify specific objects, and map their exact locations. You'll get a list of what was found and the coordinates for every item.

The results

What your AI produces

The AI processes the visual data to generate these specific outputs.

DETECTION

A list of items

The AI identifies every relevant object in the frame. You get a clean list of what's actually present in the image.

COORDINATES

Precise locations

The AI finds the exact position of each object. This allows you to know where things are located within the image bounds.

DATA

Structured visual info

The AI converts visual pixels into text-based data. You can then use this information in other parts of your workflow.

HUMAN

Manual review

You decide which detections are important. The AI flags the items, but you make the final call on how to use that data.

The workflow

What your AI does when an image arrives.

The AI handles the heavy lifting of visual processing through two specific steps.

  1. Identify items

    The AI scans the entire frame to recognize and name the objects it sees.

    detect_objects

  2. Pinpoint positions

    The AI maps the specific area where each identified object is located.

    visual_grounding

  3. Filter results

    The AI narrows down the list based on your specific requirements.

    detect_objects

  4. Verify locations

    The AI confirms the spatial relationship between different objects in the scene.

    visual_grounding

Try it

Copy these to start.

Use these prompts to begin interacting with your images.

Starting points

Replace the bracketed text with your specific image or object names.

NVIDIA Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_sRxLMKjkyXafjuH9XFRfp9b6jC4r9s5alORDsgz8/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — NVIDIA capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "nvidia-vision-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_sRxLMKjkyXafjuH9XFRfp9b6jC4r9s5alORDsgz8/mcp"
    }
  }
}
Copy into chat04

  • Detect all the objects in this photo.

  • Find the location of the red car in this image.

  • List every person visible in this frame and where they are.

  • Where is the coffee mug located in this picture?

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect NVIDIA Vision once, then ask.

One link connects the MCP to your client. Your login stays encrypted on our side, and you can find the full technical details on the Connector page.

Connect NVIDIA Vision to your AI

FAQ

How this task behaves

  • 01

    Can the AI edit the photo?

    No. The AI only identifies and locates objects; it cannot change the image itself.

  • 02

    Does the AI save my images?

    No. The AI processes the image to find objects and then moves on.

  • 03

    Can it find things it wasn't told to look for?

    Yes, if you use a general detection prompt, it will list all objects it recognizes.

  • 04

    Can the AI move objects in the frame?

    No. This MCP is for detection and grounding only.

  • 05

    How accurate are the locations?

    The AI provides coordinates based on the visual data provided in the frame.

  • More questions about NVIDIA Vision? The Connector page answers them. See everything the NVIDIA Vision Connector can do

Connect NVIDIA to Claude, Cursor, ChatGPT & more