CLI, SDKs & MCP

Three ways to interact with MUXI

MUXI provides three interfaces: the CLI for managing your server and formations, SDKs for building applications, and MCP for connecting AI-native tools like Claude Desktop and Cursor.

CLI

For: Managing MUXI

  • Deploy and manage formations
  • Configure secrets
  • Monitor logs and health
  • Server administration
muxi deploy
muxi secrets setup
muxi logs my-assistant

CLI Documentation →

SDKs

For: Building applications

  • Chat with your agents
  • Stream responses
  • Manage sessions and memory
  • Integrate into your apps
client = FormationClient(url="...")
for event in client.chat_stream({"message": "Hi"}):
    print(event.get("text"), end="")

SDK Documentation →

MCP

For: AI-native tools

  • Use from Claude Desktop or Cursor
  • 33 tools auto-generated from your API
  • Same auth as REST
  • Zero wrapper code
{
  "type": "streamable-http",
  "url": "http://localhost:7890/draft/my-formation/mcp",
  "headers": {
    "X-Muxi-Client-Key": "fmc_..."
  }
}

MCP Guide →

Which should I use?

I want to... Use
Deploy a formation CLI
Set up API keys and secrets CLI
Check formation logs CLI
Start/stop formations CLI
Build a chatbot UI SDK
Integrate agents into my app SDK
Stream responses to users SDK
Manage user sessions SDK
Use agents from Claude Desktop MCP
Use agents from Cursor MCP
Give an AI assistant access to my formation MCP

Most projects use both CLI and SDKs: CLI for deployment and ops, SDKs for application code. Add MCP when you want AI assistants to use your formation as a tool.

CLI Quick Start

Install

brew install muxi-ai/tap/muxi
curl -fsSL https://muxi.org/install | sudo bash
powershell -c "irm https://muxi.org/install | iex"

Essential Commands

# Create a new formation
muxi new formation my-assistant

# Set up secrets (API keys)
muxi secrets setup

# Run locally for development
muxi up

# Deploy to production server
muxi deploy --profile production

# View logs
muxi logs my-assistant --follow

# List all formations
muxi remote list

Full CLI Cheatsheet →

SDK Quick Start

Install

pip install muxi-client
npm install @muxi-ai/muxi-typescript
go get github.com/muxi-ai/muxi-go

Chat with Your Agent

from muxi import FormationClient

client = FormationClient(
    server_url="http://localhost:7890",
    formation_id="my-assistant",
    client_key="<your-client-key>",
    mode="draft",
)

for event in client.chat_stream({"message": "Hello!"}, user_id="user_123"):
    if event.get("type") == "text":
        print(event.get("text"), end="")
import { FormationClient } from "@muxi-ai/muxi-typescript";

const client = new FormationClient({
  serverUrl: "http://localhost:7890",
  formationId: "my-assistant",
  clientKey: "<your-client-key>",
  mode: "draft",
});

for await (const event of client.chatStream({ message: "Hello!" }, "user_123")) {
  if (typeof event.token === "string") process.stdout.write(event.token);
}
client := muxi.NewFormationClient(&muxi.FormationConfig{
    ServerURL:   "http://localhost:7890",
    FormationID: "my-assistant",
    ClientKey:   "<your-client-key>",
    Mode:        "draft",
})

stream, _ := client.ChatStream(ctx, &muxi.ChatRequest{
    Message: "Hello!",
    UserID:  "user_123",
})
for chunk := range stream {
    if token, ok := chunk.Raw["token"].(string); ok {
        fmt.Print(token)
    }
}

Full SDK Documentation →

Learn More

CLI

SDKs

Home Docs SDKs
Star on GitHub