Full Parameter Surface of MiniMax image-01
Prompt engineering, negative prompts, aspect ratios, and seeds
Exposes all underlying generation controls to the LLM with native TypeScript typing and parameter boundary checking.
Integrates the cutting-edge MiniMax image-01 text-to-image API directly into agentic coding workflows.
minimax-image-mcp is an official open source MCP tool created by Alex Santos. It connects autonomous agents to MiniMax's high-fidelity image-01 generative engine, supporting prompt styling, aspect ratios, and seed reproduction.
Handles async task polling, URL resolution, and local asset storage seamlessly.
Design choices, architectural patterns, and engineering decisions implemented for extreme reliability and developer experience.
Exposes all underlying generation controls to the LLM with native TypeScript typing and parameter boundary checking.
Manages asynchronous generation jobs by polling MiniMax task endpoints with exponential backoff and transparent error reporting.
Streams generated web images down to local disk and returns markdown image links for immediate display inside chat interfaces.
Distributed as a standalone package on the npm registry, allowing one-line configuration in any MCP-compatible client.
Real-world usage, terminal configuration, and technical integration commands.
{
"mcpServers": {
"minimax-image": {
"command": "npx",
"args": ["-y", "minimax-image-mcp"],
"env": {
"MINIMAX_API_KEY": "YOUR_API_KEY",
"MINIMAX_GROUP_ID": "YOUR_GROUP_ID"
}
}
}
}
Key clarifications regarding licensing, compatibility, deployment, and security.
On the official MiniMax platform (minimaxi.com or minimax.io).
1:1 square, 16:9 widescreen, 9:16 portrait, 4:3 classic, and 3:4 portrait.
Yes, by passing an explicit integer seed value to the generation prompt.
Currently focused on pure text-to-image generation and prompt-guided iterations.