MiniMax releases MMX-CLI: a command-line interface that offers AI brokers native entry to photographs, video, audio, music, imaginative and prescient, and search

MiniMax, the AI ​​analysis firm behind the MiniMax omnimodal mannequin stack, has launched MMX-CLI, a Node.js-based command-line interface. This exposes the complete suite of era capabilities of the MiniMax AI platform to each human builders working in terminals and AI brokers working in instruments reminiscent of Cursor, Claude Code, and OpenCode.

What drawback does MMX-CLI remedy?

Most present large-scale language mannequin (LLM)-based brokers are good at studying and writing textual content. They will purpose by paperwork, generate code, and reply to a number of directions. Nevertheless, there is no such thing as a direct path to generate media. There is no such thing as a built-in option to synthesize audio, compose music, render video, or perceive photographs with out utilizing a separate integration layer reminiscent of Mannequin Context Protocol (MCP).

Constructing these integrations sometimes requires writing customized API wrappers, configuring server-side instruments, and managing authentication individually from the agent framework you are utilizing. MMX-CLI is positioned in its place method. We expose all of those capabilities as shell instructions that brokers can name straight, the identical means builders do from the terminal. No MCP glue required.

7 modalities

MMX-CLI wraps MiniMax’s full modal stack into seven teams of generated instructions. mmx textual content, mmx picture, mmx video, mmx speech, mmx music, mmx imaginative and prescientand mmx search — Plus Help Utility (mmx auth, mmx config, mmx quota, mmx replace).

• of mmx textual content This command helps multi-turn chat, streaming output, system immediate, and JSON output modes. it accepts --model Flags to focus on particular MiniMax mannequin variants reminiscent of: MiniMax-M2.7-highspeedand MiniMax-M2.7 as default.
• of mmx picture The command generates a picture from a textual content immediate utilizing side ratio management (--aspect-ratio) and variety of batches (--n). Additionally, --subject-ref Parameters for topic references. This enables consistency of characters or objects throughout a number of photographs generated. That is helpful for workflows that require visible continuity.
• of mmx video command use MiniMax-Hailuo-2.3 Because the default mannequin, MiniMax-Hailuo-2.3-Quick Obtainable as a substitute. By default, mmx video generate Submit the job and ballot synchronously till the video is prepared. passing --async or --no-wait This habits will change. The command returns the duty ID instantly, permitting the caller to independently examine the progress. mmx video activity get --task-id. This command: --first-frame <path-or-url> Flag for picture conditional video era. The precise picture shall be used because the beginning body of the output video.
• of mmx speech This command exposes text-to-speech (TTS) synthesis, velocity management, quantity and pitch adjustment, and subtitle timing information output with over 30 accessible voices. --subtitlesassist for streaming playback through a pipe to a media participant. The default mannequin is speech-2.8-hdand speech-2.6 and speech-02 in its place. The enter restrict is 10,000 characters.
• of mmx music command, music-2.5 The mannequin generates music from textual content prompts utilizing fine-grained composition controls, together with: --vocals (for instance "heat male baritone"), --genre, --mood, --instruments, --tempo, --bpm, --keyand --structure. of --instrumental flag produces music with out vocals. Ann --aigc-watermark Flags will also be used to embed watermarks of AI-generated content material into the output audio.
• mmx imaginative and prescient Course of picture understanding through Imaginative and prescient Language Mannequin (VLM). Accepts a neighborhood file path or distant URL (native recordsdata are mechanically Base64 encoded), or a beforehand uploaded MiniMax file ID. a --prompt Flags permit you to ask particular questions on photographs. The default immediate is "Describe the picture."
• mmx search Run net search queries by MiniMax’s proprietary search infrastructure and return leads to textual content or JSON format.

expertise structure

MMX-CLI is written nearly completely in TypeScript (99.8% TS) with Strict mode enabled, makes use of Bun because the native runtime for improvement and testing, and is distributed on npm for compatibility with Node.js 18+ environments. Configuration schema validation makes use of Zod and determination follows outlined priorities (CLI Flags → Atmosphere Variables → ~/.mmx/config.json → Default—Simplifies deployment in containerized or CI environments. Twin-region assist is constructed into the API shopper tier to assist world customers. api.minimax.io and CN customers, api.minimaxi.commight be switched through mmx config set --key area --value cn.

Necessary factors

• MMX-CLI is MiniMax’s official open command line interface, giving AI brokers native entry to seven generative modalities: textual content, photographs, video, audio, music, imaginative and prescient, and search, with out requiring MCP integration.
• AI brokers working in instruments like Cursor, Claude Code, and OpenCode might be arrange with two instructions and one pure language instruction, after which the agent learns the whole command interface by itself from the bundled SKILL.md doc.
• The CLI is designed to be used by applications and brokers, with devoted flags for non-interactive execution, clear stdout/stderr separation for protected pipes, structured exit codes for error dealing with, and a schema export characteristic that permits agent frameworks to register mmx instructions as JSON software definitions.
• For AI builders already constructing agent-based methods, integrating picture, video, audio, music, imaginative and prescient, and search era right into a single, well-documented CLI that brokers can be taught and function on their very own considerably lowers the mixing barrier.

Please examine Click here for the report. Please be at liberty to observe us too Twitter Do not forget to hitch us 130,000+ ML subreddits and subscribe our newsletter. grasp on! Are you on telegram? You can now also participate by telegram.

Have to companion with us to advertise your GitHub repository, Hug Face Web page, product launch, webinar, and so forth.? connect with us