VisionCore-MCP 👁️🤖

Give Claude eyes. Local. Private. Real-time.

VisionCore-MCP is an easy-to-use bridge that allows Claude (Desktop) to see through your webcam. It uses local AI to "reason" about what it sees, allowing you to ask questions about your physical surroundings without your data ever leaving your computer.

🌟 Features

• Ask Claude Anything: "What am I holding?", "Is my coffee still hot?", "Did someone walk past my door?"
• Temporal Memory: Claude remembers what happened in your room over the last 5 minutes.
• 100% Private: Everything runs locally via Ollama. No cloud vision APIs, no subscription fees.

---

🛠️ Easy Setup (3 Steps)

1. Install the "Brain" (Ollama)

Download and install Ollama. Once installed, open your terminal and run:

ollama pull llava

2. Set Up the "Eyes" (Python)

Ensure you have Python 3.12+ installed.

1. Open this folder in your terminal.
2. Run the setup command:

   py -m venv venv
   .\venv\Scripts\activate
   pip install -r requirements.txt

3. Connect to Claude

1. Open Claude Desktop settings: %APPDATA%\Claude\claude_desktop_config.json
2. Add this to your mcpServers list (adjusting the path to your folder):

{
  "mcpServers": {
    "vision-core": {
      "command": "C:\\Projects\\visiioncore-mcp\\venv\\Scripts\\python.exe",
      "args": ["C:\\Projects\\visiioncore-mcp\\server.py"]
    }
  }
}

3. Restart Claude Desktop.

---

🚀 How to Use

Once the 🔨 icon appears in Claude, try asking:

• "Check the environment: What is in front of my camera?"
• "Look at my desk. Does it look cluttered?"
• "List available cameras" (If you have more than one webcam)

---

🔍 Troubleshooting

• No Image?: Make sure your webcam isn't being used by Zoom, Teams, or another app.
• Slow?: VisionCore runs best on a GPU. If you don't have one, Ollama will use your CPU, which might take a few seconds per query.
• Still "Urn"?: Lighting is key! Make sure your object is well-lit and held clearly in front of the camera.

---

📚 Advanced Documentation

• Architecture Deep Dive
• Technical Implementation Details