vision-mcp-server
Verified Safeby hiroki-yokoyama
Overview
A Model Context Protocol (MCP) server for local CPU-based vision language model inference using GGUF models via llama-cpp-python, designed to run as a Windows resident process and analyze images.
Installation
scripts\run_server.ps1Environment Variables
- HF_ENDPOINT
- HF_TOKEN
Security Notes
The server primarily operates with local file paths for images and models, relying on PIL for image processing and llama-cpp-python for LLM inference. No direct 'eval' or execution of arbitrary code from user input was found. The dynamic loading of chat handlers uses a controlled dictionary, mitigating risks. Main risks involve running potentially malicious GGUF models or providing large/malformed image files, which are inherent to the use case.
Similar Servers
osaurus
Osaurus is an AI edge runtime for macOS, enabling users to run local and cloud AI models, orchestrate tools via the Model Context Protocol (MCP), and power AI applications and workflows on Apple Silicon.
luma-mcp
Provides multi-model vision understanding capabilities to AI assistants that lack native image understanding.
ollama-mcp-server
Provides a self-contained Model Context Protocol (MCP) server for local Ollama management, enabling features like listing models, chatting, server control, and intelligent model recommendations.
polybrain-mcp
Connects AI agents to multiple LLM models, providing conversation history management and model switching capabilities.