Integrates Xiaomi's MiMo API to use MiMo models (e.g., MiMo V2.5 Pro) as reference agents in a mixture-of-agents discussion, where multiple models analyze a query in parallel and a primary model aggregates the responses to generate a final answer.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MoA MCP Server discuss: pros and cons of electric vehicles"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Mixture-of-Agents MCP Server — 并行调用多个参考模型讨论问题，再由总统筹模型综合输出最佳答案。
你的问题
↓
[DeepSeek V4 Flash] [MiMo V2.5 Pro] ← 参考模型并行分析
↓ ↓
[DeepSeek V4 Pro] ← 总统筹综合输出
↓
最终答案 + 参考意见原文Related MCP server: brainstorm-mcp
变量 | 必填 | 说明 |
| ✅ | DeepSeek API Key |
| ✅ | 小米 MiMo API Key |
| 可选 | 默认 |
| 可选 | 参考模型1，默认 |
| 可选 | 默认 |
| 可选 | 参考模型2，默认 |
| 可选 | 总统筹模型，默认 |
在 Proma 的 mcp.json 中添加：
{
"moa": {
"command": "node",
"args": ["/path/to/moa-mcp-server/server.js"],
"env": {
"DEEPSEEK_API_KEY": "sk-xxx",
"MIMO_API_KEY": "sk-xxx",
"FLASH_MODEL": "deepseek-chat",
"MIMO_MODEL": "mimo-2.5-pro",
"AGGREGATOR_MODEL": "deepseek-v4-pro"
},
"type": "stdio",
"enabled": true
}
}支持任何兼容 MCP 协议的客户端（Claude Desktop、VS Code 等），按对应平台配置 stdio MCP server 即可。
在 Agent 会话中调用 moa_discuss 工具：
query（必填）：需要讨论的问题
context（可选）：背景信息
将问题并行发送给 DeepSeek V4 Flash 和 MiMo V2.5 Pro
两个参考模型独立分析，返回各自意见
总统筹模型 (DeepSeek V4 Pro) 综合所有参考意见，输出最终答案
返回结果包含最终答案 + 各参考模型的原始意见原文
This server cannot be deployed
Multi-model AI debates: GPT-4o, Claude, Gemini & 200+ models discuss, then synthesize insight.
Agent-to-agent reasoning-as-a-service: chain-of-thought, analysis, and decision support.
Multi-expert decision intelligence with transparent synthesis and auditable workflows.
Convene a panel of expert AI personas to debate any decision from every side.
