Running MCP 40 Serverless TextGen Hub ♨ 40 Now supports inference providers and multimodal. No GPU req.
