# Costruire un client MCP personalizzato

Se sta costruendo la propria applicazione LLM, può connettersi al server MCP
di Synapse direttamente utilizzando l'SDK MCP ufficiale. In questo modo la
sua app avrà accesso a tutti i 79 strumenti di Synapse.

## SDK

| Linguaggio | Pacchetto |
|----------|---------|
| TypeScript/JavaScript | `@modelcontextprotocol/sdk` |
| Python | `mcp` |

## Esempio TypeScript

### Installazione

```bash
npm install @modelcontextprotocol/sdk
```

### Connessione tramite stdio

```typescript
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";

const transport = new StdioClientTransport({
  command: "npx",
  args: ["-y", "synapse-mcp-api@latest"],
  env: {
    SYNAPSE_MIND_KEY: process.env.SYNAPSE_MIND_KEY!,
    SYNAPSE_URL: "https://synapse.schaefer.zone",
  },
});

const client = new Client(
  { name: "my-app", version: "1.0.0" },
  { capabilities: {} }
);

await client.connect(transport);

// List all available tools
const { tools } = await client.listTools();
console.log(`Available tools: ${tools.length}`);
for (const tool of tools) {
  console.log(`- ${tool.name}: ${tool.description}`);
}

// Call a tool
const result = await client.callTool({
  name: "memory_recall",
  arguments: {},
});
console.log(result.content);

// Store a memory
await client.callTool({
  name: "memory_store",
  arguments: {
    category: "fact",
    key: "custom_client_test",
    content: "Built a custom MCP client",
    tags: ["test", "mcp"],
    priority: "normal",
  },
});

await client.close();
```

### Connessione tramite HTTP/SSE (remoto)

```typescript
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { SSEClientTransport } from "@modelcontextprotocol/sdk/client/sse.js";

const transport = new SSEClientTransport(
  new URL("https://synapse-mcp.schaefer.zone/sse"),
  {
    requestInit: {
      headers: {
        Authorization: `Bearer ${process.env.SYNAPSE_MIND_KEY}`,
      },
    },
  }
);

const client = new Client(
  { name: "my-app", version: "1.0.0" },
  { capabilities: {} }
);

await client.connect(transport);
// ... use as above
```

## Esempio Python

### Installazione

```bash
pip install mcp
```

### Connessione tramite stdio

```python
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client

server_params = StdioServerParameters(
    command="npx",
    args=["-y", "synapse-mcp-api@latest"],
    env={
        "SYNAPSE_MIND_KEY": "mk_YOUR_KEY",
        "SYNAPSE_URL": "https://synapse.schaefer.zone",
    },
)

async with stdio_client(server_params) as (read, write):
    async with ClientSession(read, write) as session:
        await session.initialize()

        # List tools
        tools = await session.list_tools()
        print(f"Available tools: {len(tools.tools)}")

        # Call a tool
        result = await session.call_tool("memory_recall", {})
        print(result.content)

        # Store a memory
        await session.call_tool("memory_store", {
            "category": "fact",
            "key": "python_client_test",
            "content": "Built a Python MCP client",
            "tags": ["test", "mcp", "python"],
            "priority": "normal",
        })
```

## Profili degli strumenti

Quando si connette, può richiedere un profilo di strumenti specifico tramite
l'header `Mcp-Tool-Profile` (HTTP/SSE) o la variabile d'ambiente `MCP_PROFILE`
(stdio):

```typescript
// stdio: set env var
env: {
  SYNAPSE_MIND_KEY: "mk_...",
  MCP_PROFILE: "minimal",  // 8 tools instead of 119
}

// HTTP/SSE: set header
requestInit: {
  headers: {
    Authorization: "Bearer mk_...",
    "Mcp-Tool-Profile": "minimal",
  },
}
```

## Gestione degli errori

```typescript
try {
  const result = await client.callTool({ name: "memory_recall", arguments: {} });
  if (result.isError) {
    console.error("Tool error:", result.content);
  } else {
    console.log("Success:", result.content);
  }
} catch (err) {
  console.error("MCP error:", err);
}
```

## Casi d'uso

- **Assistenti AI personalizzati** — costruisca il proprio agente con memoria persistente
- **Automazione dei flussi di lavoro** — concatena strumenti Synapse in flussi personalizzati
- **Pipeline di dati** — estrae memorie, trasforma, carica altrove
- **Dashboard di monitoraggio** — visualizza statistiche delle memorie, cronologia chat, attività

## Prossimi passi

- [Specifica MCP](https://spec.modelcontextprotocol.io)
- [Repository Synapse MCP](https://gitlab.com/schaefer-services/synapse-mcp)
- [Panoramica API](/docs/api/overview)