Skip to main content
For Token Plan users, no coding is required to unlock MiniMax multimodal capabilities in your AI agent: video generation, speech synthesis, music creation, coding, and more. You can call these capabilities directly in assistants like OpenClaw and Claude Code. If you still prefer direct API integration, see the API Documentation.

Install the CLI

Copy the prompt below to your AI agent (OpenClaw, Claude Code, Cursor, MaxClaw, AutoClaw, KimiClaw, TRAE, OpenCode, etc.). It will guide you to add your API Key and finish the setup:
Help me install MiniMax CLI: https://github.com/MiniMax-AI/cli

Use the CLI

TextUse minimax to write a 4-line poem about AI
VideoGenerate a video: at sunset, a cat sits by the window looking into the distance
MusicGenerate an upbeat jazz song about a summer beach
SpeechRead in a gentle female voice: Welcome to MiniMax Token Plan. After subscribing, your AI agent can generate video, music, speech, and images with full multimodal capability.
ImageGenerate a cyberpunk city night scene in 16:9Generated files are saved in minimax-output/ under your current directory. If you’re using an agent, it’s recommended to have it display the generated media directly in its output.

CLI Dashboard

Run mmx in your terminal to open the CLI panel and quickly discover the main commands, flags, and usage info.

MMX-CLI panel overview
  • resources: available resource types
  • flags: supported options for commands
  • usage: remaining quota and usage overview
  • help: entry points for documentation

Capability overview

MMX-CLI provides a single command-line entry point across text, image, video, speech, music, vision understanding, and web search:
CapabilityDescription
Textmulti-turn chat, streaming output, system prompts, JSON output
Imagetext-to-image, aspect ratio controls, batch generation
Videoasync generation, task status, downloading
Speechtext-to-speech (TTS), multiple voices, streaming
Musictext-to-music, with-lyrics and instrumental modes
Visionimage understanding from local files, URLs, or file IDs
Searchbuilt-in web search

Usage by modality

For modality-specific Token Plan quotas, see: Token Plan Pricing