MiniMax Launches MMX-CLI: A Command-Line Interface for AI Agents

MiniMax, an AI research company, has unveiled MMX-CLI, a Node.js-based command-line interface that grants access to the full suite of generative capabilities of the MiniMax AI platform for both human developers working in a terminal and AI agents operating in tools like Cursor, Claude Code, and OpenCode.

Currently, most large language model (LLM)-based agents excel at reading and writing text, reasoning over documents, generating code, and responding to multi-turn instructions. However, they lack a direct pathway for media generation, which typically requires a separate integration layer such as the Model Context Protocol (MCP). Building these integrations often involves writing custom API wrappers and configuring server-side tools, complicating the process.

MMX-CLI serves as an alternative approach, allowing all these capabilities to be exposed as shell commands that agents can invoke directly, just as developers do from a terminal, without the need for MCP integration. The interface wraps MiniMax’s full-modal stack into seven generative command groups: mmx text, mmx image, mmx video, mmx speech, mmx music, mmx vision, and mmx search, along with supporting utilities.

The mmx text command supports multi-turn chat, streaming output, and JSON mode. The mmx image command generates images from text prompts with controls for aspect ratio and batch count. The mmx video command defaults to MiniMax-Hailuo-2.3 and allows for video generation using a specified image as the opening frame.

The mmx speech command provides text-to-speech synthesis with over 30 available voices and controls for speed and volume. The mmx music command generates music from text prompts with fine-grained compositional controls. The mmx vision command handles image understanding via a vision-language model, while mmx search executes web search queries through MiniMax’s infrastructure.

MMX-CLI is primarily written in TypeScript and utilizes Bun for development and testing, ensuring compatibility with Node.js 18+. The interface is designed for programmatic and agent use, significantly lowering the integration barrier for developers building agent-based systems.