Running MCP 44 Serverless TextGen Hub ♨ 44 Now supports inference providers and multimodal. No GPU req.
