> ## Documentation Index
> Fetch the complete documentation index at: https://docs.neosantara.xyz/llms.txt
> Use this file to discover all available pages before exploring further.

# Server Tools & Integrasi MCP

> Eksekusi tools server-side dan integrasi remote MCP server menggunakan OpenAI SDK.

Responses API (`/v1/responses`) mendukung dua jenis tools: pemanggilan fungsi standar di sisi klien (*client-side function tools*) dan tools yang didelegasikan langsung oleh gateway ke server remote melalui **Model Context Protocol (MCP)**.

## Integrasi Server MCP Remote

Dengan menambahkan konfigurasi server MCP ke dalam array `tools`, Neosantara bertindak sebagai klien MCP dan mengeksekusi alat secara langsung di lapisan gateway tanpa melibatkan kode orkestrasi di sisi klien.

<CodeGroup>
  ```python Python (OpenAI SDK) icon="python" theme={"theme":{"light":"ayu-dark","dark":"catppuccin-latte"}}
  from openai import OpenAI
  import os

  client = OpenAI(
      base_url="https://api.neosantara.xyz/v1",
      api_key=os.environ["NEOSANTARA_API_KEY"]
  )

  # Sertakan server MCP remote di dalam array tools
  response = client.responses.create(
      model="deepseek-v4.1-flash",
      input="Periksa daftar isu bug prioritas tinggi di repositori kami.",
      tools=[
          {
              "type": "mcp",
              "server_label": "github",
              "server_url": "https://mcp.github-service.example.com/mcp",
              "headers": {
                  "Authorization": "Bearer gh_mcp_token_secret"
              },
              "allowed_tools": ["list_issues", "get_issue"]
          }
      ]
  )

  print(response.output_text)
  ```

  ```javascript TypeScript (Node.js) icon="js" theme={"theme":{"light":"ayu-dark","dark":"catppuccin-latte"}}
  import OpenAI from "openai";

  const client = new OpenAI({
    baseURL: "https://api.neosantara.xyz/v1",
    apiKey: process.env.NEOSANTARA_API_KEY,
  });

  const response = await client.responses.create({
    model: "deepseek-v4.1-flash",
    input: "Cek inventaris stok barang untuk SKU-9821.",
    tools: [
      {
        type: "mcp",
        server_label: "inventory",
        server_url: "https://mcp.warehouse.example.com/mcp",
        headers: {
          Authorization: "Bearer warehouse_secret_token",
        },
      },
    ],
  });

  console.log(response.output_text);
  ```
</CodeGroup>

## Deklarasi Fungsi Client-Side

Jika Anda ingin model menentukan parameter dan membiarkan aplikasi klien Anda yang menjalankan logika lokalnya, deklarasikan skema fungsi standar JSON.

<CodeGroup>
  ```python Python (OpenAI SDK) icon="python" theme={"theme":{"light":"ayu-dark","dark":"catppuccin-latte"}}
  from openai import OpenAI
  import os

  client = OpenAI(
      base_url="https://api.neosantara.xyz/v1",
      api_key=os.environ["NEOSANTARA_API_KEY"]
  )

  # Deklarasi fungsi kustom untuk dievaluasi oleh model
  response = client.responses.create(
      model="deepseek-v4.1-flash",
      input="Berapa perkiraan cuaca di Bandung hari ini?",
      tools=[
          {
              "type": "function",
              "function": {
                  "name": "get_weather",
                  "description": "Ambil cuaca terkini berdasarkan kota",
                  "parameters": {
                      "type": "object",
                      "properties": {
                          "city": {"type": "string"}
                      },
                      "required": ["city"]
                  }
              }
          }
      ]
  )

  print(response.output_text)
  ```

  ```javascript TypeScript (Node.js) icon="js" theme={"theme":{"light":"ayu-dark","dark":"catppuccin-latte"}}
  import OpenAI from "openai";

  const client = new OpenAI({
    baseURL: "https://api.neosantara.xyz/v1",
    apiKey: process.env.NEOSANTARA_API_KEY,
  });

  const response = await client.responses.create({
    model: "deepseek-v4.1-flash",
    input: "Berapa perkiraan cuaca di Bandung hari ini?",
    tools: [
      {
        type: "function",
        function: {
          name: "get_weather",
          description: "Ambil cuaca terkini berdasarkan kota",
          parameters: {
            type: "object",
            properties: {
              city: { type: "string" },
            },
            required: ["city"],
          },
        },
      },
    ],
  });

  console.log(response.output_text);
  ```
</CodeGroup>

## Keunggulan Server MCP di Responses API

Pada Chat Completions biasa, ketika model AI memutuskan memanggil tool, klien harus menangkap `tool_calls`, mengeksekusi kode sendiri di server lokal, lalu mengirimkan balik hasilnya ke gateway.

Pada Responses API dengan tool tipe `mcp`:

1. **Gateway Mengeksekusi Otomatis:** Neosantara menghubungi server MCP Anda secara langsung.
2. **Klien Cukup Menunggu Respons:** Klien tidak perlu menulis loop orkestrasi pemanggilan fungsi.
3. **Kompatibel dengan Mode Asinkron:** Sangat cocok dipadukan dengan `background=True` untuk proses data intensif.

## Langkah Selanjutnya

| Kebutuhan                     | Panduan                                                       |
| :---------------------------- | :------------------------------------------------------------ |
| Panduan MCP Connector Lengkap | [MCP Connector](/id/agents/mcp-connector)                     |
| Tugas Latar Belakang          | [Background Tasks](/id/gateway/responses-api/background-jobs) |
| Simpan State Percakapan       | [Percakapan & State](/id/gateway/responses-api/conversations) |


## Related topics

- [OpenResponses API](/id/gateway/responses-api.md)
- [Function Calling (Tools)](/id/gateway/chat-completions/tool-calling.md)
- [MCP Connector](/id/agents/mcp-connector.md)