shimmy (MCP Server)
GitHub⚡ Pure-Rust WebGPU inference engine — OpenAI-API compatible, GGUF native, runs on any GPU. No Python. No llama.cpp. Single binary.
Claude Desktop Configuration
Add to ~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"shimmy": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-shimmy"
]
}
}
}