Add MiniMax LLM Provider Support

[285480731]

Hi, I love PicoClaw! It would be great if you could add support for MiniMax API.

Supported models:

• MiniMax-Text-01
• MiniMax-Vision-01

This would allow more users to access affordable LLM services.

Thank you!

[Tuzfucius]

你好！你可以通过 MiniMax 提供的 OpenAI 兼容接口来接入模型。

请编辑配置文件 ~/.picoclaw/config.json，参考以下配置进行修改：

{
  "agents": {
    "defaults": {
      "workspace": "~/.picoclaw/workspace",
      "restrict_to_workspace": true,
      "provider": "openai",
      "model": "MiniMax-M2.1",
      "max_tokens": 5000,
      "temperature": 0.7,
      "max_tool_iterations": 20
    }
  },
  "providers": {
    "openai": {
      "api_key": "YOUR_MINIMAX_API_KEY",
      "api_base": "https://api.minimaxi.com/v1"
    }
  }
}

api_base可能因为地区差异而不同，请参考服务商的官方文档
但是据我所知 Minimax 是通过 mcp 服务来进行图像识别的，需要额外安装

[happyTonakai]

providers are DEPRECATED: Use model_list instead. This will be removed in a future version.

[nikolasdehor]

As @Tuzfucius mentioned, this already works via the OpenAI-compatible provider. Updated config using the current model_list format (since providers is deprecated):

{
  "model_list": [
    {
      "model_name": "minimax",
      "provider": "openai",
      "model": "MiniMax-Text-01",
      "api_key": "YOUR_MINIMAX_API_KEY",
      "api_base": "https://api.minimaxi.com/v1"
    }
  ]
}

For first-class minimax/ prefix support, the change in pkg/providers/factory_provider.go would be minimal — add "minimax" to the OpenAI-compatible case and set the default base URL. This is the same pattern used for openrouter, together, groq, etc.

Related: #729 is requesting the same thing. These two issues could be consolidated.

[keithy]

Docs at https://platform.minimax.io/docs/guides/text-ai-coding-tools say:
Enter the MiniMax base URL (for international users, use https://api.minimax.io/v1; for users in China, use https://api.minimaxi.com/v1)

[jhao0413]

你好！你可以通过 MiniMax 提供的 OpenAI 兼容接口来接入模型。

请编辑配置文件 ~/.picoclaw/config.json，参考以下配置进行修改：

{
"agents": {
"defaults": {
"workspace": "~/.picoclaw/workspace",
"restrict_to_workspace": true,
"provider": "openai",
"model": "MiniMax-M2.1",
"max_tokens": 5000,
"temperature": 0.7,
"max_tool_iterations": 20
}
},
"providers": {
"openai": {
"api_key": "YOUR_MINIMAX_API_KEY",
"api_base": "https://api.minimaxi.com/v1"
}
}
}
api_base可能因为地区差异而不同，请参考服务商的官方文档 但是据我所知 Minimax 是通过 mcp 服务来进行图像识别的，需要额外安装

请问需要安装什么mcp来支持图像识别呢

[jhao0413]

找了一下，官方是有提供mcp的，但是目前0.2.0的版本好像不支持mcp，配置后没有生效