Kolosal Vision

Provides AI-powered image analysis and OCR via the Kolosal Vision API for MCP-compatible clients.

typescript

0

GitHub Stars

typescript

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "madebyaris-kolosal-vision-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "kolosal-vision-mcp"
      ],
      "env": {
        "KOLOSAL_API_KEY": "your_api_key_here"
      }
    }
  }
}

You can add Kolosal Vision as an MCP server to your client to enable AI-powered image analysis and OCR directly from images, URLs, local files, or base64 data. It returns structured observations and supports common image formats, making it easy to ask questions like what objects appear, what text is visible, or what the scene contains.

How to use

To use Kolosal Vision as an MCP server, configure your MCP client to point at one or more stdio-based instances that run the Kolosal Vision MCP. You can run the server locally using npx or a global installation, then reference images from URLs, your filesystem, or base64 data in your prompts. Typical usage involves starting the server and then asking questions like “Describe this image” or “Extract any visible text.” You can combine image sources with natural language queries to get structured observations.

How to install

Prerequisites: Node.js 18+ and npm or yarn.

# Using npx (recommended):
 npx kolosal-vision-mcp

# Global installation:
 npm install -g kolosal-vision-mcp

# Local installation:
 npm install kolosal-vision-mcp

Additional setup and usage notes

Configure your MCP client with one or more server entries. The following examples show how to connect via stdio for Cursor IDE and Claude Desktop, using an API key obtained from Kolosal AI.

{
  "mcpServers": {
    "kolosal-vision": {
      "command": "npx",
      "args": ["-y", "kolosal-vision-mcp"],
      "env": {
        "KOLOSAL_API_KEY": "your_api_key_here"
      }
    }
  }
}

How to use with Claude Desktop and Cursor IDE

With Cursor IDE or Claude Desktop, add a server entry that runs the MCP and provides your API key via environment variables. You will reference images in prompts and receive structured results.

{
  "mcpServers": {
    "kolosal-vision": {
      "command": "npx",
      "args": ["-y", "kolosal-vision-mcp"],
      "env": {
        "KOLOSAL_API_KEY": "your_api_key_here"
      }
    }
  }
}

Usage prompts and expected responses

Examples of prompts you can ask the MCP server include: “What objects are in this image?”, “Describe the scene in detail”, “Extract any visible text (OCR)”, “Are there any people? What are they doing?”, “What brand logos are visible?”. The server returns a structured analysis with key observations.

Response format

You will receive a structured analysis plus a list of key observations to help you quickly grasp the important details of the image.

Security and API key handling

Keep your KOLOSAL_API_KEY secure. Do not share it in public prompts or logs. Store it in a secure environment variable for your MCP client configuration.

Available tools

analyze_image

Analyzes an image given an image source and a description prompt, returning a structured set of observations and details.