Dokumentasi API

play_arrow Quick Start Python

OpenAI SDK Python, cuma ganti base_url.

from openai import OpenAI

client = OpenAI(
    api_key="divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
    base_url="https://divproxy.my.id/v1"
)

response = client.chat.completions.create(
    model="qwen3.6-plus",
    messages=[{"role": "user", "content": "Halo!"}]
)
print(response.choices[0].message.content)

terminal cURL

curl https://divproxy.my.id/v1/chat/completions \
  -H "Authorization: Bearer divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "messages": [{"role": "user", "content": "Halo!"}]
  }'

javascript Node.js

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  baseURL: "https://divproxy.my.id/v1",
});

const response = await client.chat.completions.create({
  model: "qwen3.6-plus",
  messages: [{ role: "user", content: "Halo!" }],
});

console.log(response.choices[0].message.content);

terminal Claude Code

Pakai base URL tanpa /v1. Claude Code akan call /v1/messages sendiri.

# PowerShell
$env:ANTHROPIC_BASE_URL="https://divproxy.my.id/anthropic"
$env:ANTHROPIC_API_KEY="divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
$env:ANTHROPIC_CUSTOM_MODEL_OPTION="qwen3.6-plus"
$env:ANTHROPIC_MODEL="qwen3.6-plus"
$env:CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY="1"
claude

# macOS / Linux
export ANTHROPIC_BASE_URL="https://divproxy.my.id/anthropic"
export ANTHROPIC_API_KEY="divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
export ANTHROPIC_CUSTOM_MODEL_OPTION="qwen3.6-plus"
export ANTHROPIC_MODEL="qwen3.6-plus"
export CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY="1"
claude

extension Kilo Code

VS Code AI coding assistant. Sambungkan via provider OpenAI Compatible.

Buka VS Code, klik ikon Kilo Code di sidebar
Klik Settings -> API Provider
Pilih: OpenAI Compatible
Isi konfigurasi di bawah, klik Save

Base URL  : https://divproxy.my.id/v1
API Key   : divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Model     : qwen3.6-plus

Untuk thinking dan tool use lebih native, pilih provider Anthropic:

Base URL  : https://divproxy.my.id/anthropic
API Key   : divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Model     : qwen3.6-plus

code OpenCode

Terminal AI coding assistant. Pakai env var atau file config.

OPENAI_API_KEY="divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" \
OPENAI_BASE_URL="https://divproxy.my.id/v1" \
opencode

Atau tambahkan ke ~/.config/opencode/config.json:

{
  "provider": "openai",
  "model": "qwen3.6-plus",
  "baseURL": "https://divproxy.my.id/v1",
  "apiKey": "divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}

article Continue

VS Code dan JetBrains extension. Edit file ~/.continue/config.json:

{
  "models": [
    {
      "title": "DivProxy Qwen3.6-Plus",
      "provider": "openai",
      "model": "qwen3.6-plus",
      "apiKey": "divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
      "apiBase": "https://divproxy.my.id/v1"
    }
  ]
}

Simpan, reload VS Code, model muncul di dropdown Continue.

smart_toy Cline / Roo Code

VS Code extension agentic coding. Support OpenAI-compatible dan Anthropic.

Via OpenAI Compatible:

Base URL  : https://divproxy.my.id/v1
API Key   : divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Model ID  : qwen3.6-plus

Via Anthropic (recommended untuk agentic + thinking):

Base URL  : https://divproxy.my.id/anthropic
API Key   : divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Model ID  : qwen3.6-plus

Mode Anthropic dukung extended thinking, tool use, vision native.

terminal Aider

Terminal AI coding assistant. Pakai flag --openai-api-base dan --model:

aider \
  --openai-api-key "divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" \
  --openai-api-base "https://divproxy.my.id/v1" \
  --model "openai/qwen3.6-plus"

Atau lewat env, tambahkan ke .bashrc atau .zshrc:

export OPENAI_API_KEY="divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
export OPENAI_API_BASE="https://divproxy.my.id/v1"

aider --model openai/qwen3.6-plus

forum Anthropic Messages

curl https://divproxy.my.id/anthropic/v1/messages \
  -H "x-api-key: divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-plus",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Halo!"}]
  }'

php PHP

$payload = [
  "model" => "qwen3.6-plus",
  "messages" => [["role" => "user", "content" => "Halo!"]],
];

$ch = curl_init("https://divproxy.my.id/v1/chat/completions");
curl_setopt_array($ch, [
  CURLOPT_POST => true,
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer divproxy-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
    "Content-Type: application/json",
  ],
  CURLOPT_POSTFIELDS => json_encode($payload),
  CURLOPT_RETURNTRANSFER => true,
]);

echo curl_exec($ch);

stream Streaming

Set stream: true, response keluar token demi token via Server-Sent Events.

{
  "model": "qwen3.6-plus",
  "stream": true,
  "enable_thinking": true,
  "messages": [{"role": "user", "content": "Tulis pantun pendek"}]
}

image Vision

Penting: untuk saat ini model vision adalah qwen3.6-plus dan qwen3.6-plus-2026-04-02. Model lain text-only - kalau dikirim gambar akan ditolak. Bisa kirim gambar via URL atau base64. Bisa digabung thinking mode.

response = client.chat.completions.create(
    model="qwen3.6-plus",
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "Jelaskan isi gambar ini."},
            {"type": "image_url", "image_url": {"url": "https://example.com/image.jpg"}}
        ]
    }],
    extra_body={"enable_thinking": True}
)

smart_toy Model tersedia

Semua model bisa diakses via parameter "model" di request body. Contoh: "model": "qwen3.7-max".

qwen3.7-max

Qwen

Flagship reasoning. Text-only.
Versi maximum Qwen 3.7 untuk problem solving kompleks dan agentic coding tingkat lanjut. Tidak menerima input gambar.

Text-only Extended Thinking 1M Context Agentic

qwen3.6-plus

Qwen

Default model. Vision support.
Model utama untuk chat, vision, thinking, dan tool use. 1M context window. Menerima input gambar.

Vision Thinking 1M Context Tool Use

deepseek-v4-pro

DeepSeek

Reasoning + coding flagship.
DeepSeek V4 Pro untuk reasoning panjang, math, dan coding. Cocok buat refactor besar dan analisis algoritma. Text-only.

Reasoning Coding Math Text-only

deepseek-v4-flash

DeepSeek

Cepat untuk request volume tinggi.
Versi flash DeepSeek V4. Latensi rendah dengan trade-off kualitas, pas buat chatbot dan automation harian. Text-only.

Low Latency High Throughput Text-only

glm-5.1

Zhipu AI

General reasoning dari Zhipu.
GLM 5.1 dari Zhipu AI. Model general-purpose untuk reasoning, chat, dan text generation. Text-only.

Reasoning Text-only General

glm-5.2

Zhipu AI

Advanced reasoning dari Zhipu.
GLM 5.2 flagship dari Zhipu AI. Model reasoning terbaru dengan peningkatan performa untuk analisis logis, matematika, dan programming. Text-only.

Reasoning Programming Text-only

kimi-k2.5

Moonshot AI

Long context specialist.
Kimi K2.5 dari Moonshot AI. Kuat di pemrosesan dokumen panjang dan analisis teks. Text-only.

Long Context Text-only Analysis

kimi-k2.7-code

Moonshot AI

Next-gen long context & coding & reasoning.
Kimi K2.7 Code flagship dari Moonshot AI. Performa logika, pemrosesan dokumen panjang, dan coding assistant yang lebih handal. Text-only.

Reasoning Long Context Text-only

mimo-v2.5-pro

XiaomiMiMo

Limited Edition.
MiMo V2.5 Pro via XiaomiMiMo. Model Limited Edition; request tetap memotong kuota DivProxy. Kalau provider sedang rate-limit, tunggu cooldown/tersedia lagi. Text-only.

Limited Edition Text-only Kuota Terhitung

qwen3.7-max-2026-05-20

Qwen

Snapshot stable build 20 Mei 2026.
Versi pinned Qwen 3.7 Max. Hasil reproducible untuk benchmark atau testing. Text-only.

Pinned Text-only Thinking

qwen3.6-plus-2026-04-02

Qwen

Snapshot stable build 2 April 2026. Vision support.
Versi pinned Qwen 3.6 Plus dengan vision. Hasil reproducible untuk testing. Menerima input gambar.

Pinned Vision Thinking

qwen3.5-plus

Qwen

Stable production workhorse.
Qwen 3.5 Plus untuk workload production yang butuh stabilitas. Text-only, latensi rendah.

Stable Text-only Production

qwen3.5-plus-2026-02-15

Qwen

Snapshot stable build 15 Februari 2026.
Versi pinned Qwen 3.5 Plus. Hasil reproducible untuk testing. Text-only.

Pinned Text-only

qwen3-max-2025-09-23

Qwen

Snapshot stable build 23 September 2025.
Versi pinned Qwen 3 Max. Legacy model dengan hasil reproducible. Text-only.

Pinned Text-only Legacy

error Error umum

Code	Arti	Solusi
401	API key salah atau tidak ada	Pakai header `Authorization: Bearer divproxy-...`
402	Paket sudah expired	Beli ulang atau upgrade paket di console
429	Quota atau rate limit habis	Tunggu reset atau upgrade paket
503	Maintenance	Coba lagi setelah admin selesai maintenance

DivProxy API Reference

play_arrow Quick Start Python

terminal cURL

javascript Node.js

terminal Claude Code

extension Kilo Code

code OpenCode

article Continue

smart_toy Cline / Roo Code

terminal Aider

forum Anthropic Messages

php PHP

stream Streaming

image Vision

smart_toy Model tersedia

qwen3.7-max

qwen3.6-plus

deepseek-v4-pro

deepseek-v4-flash

glm-5.1

glm-5.2

kimi-k2.5

kimi-k2.7-code

mimo-v2.5-pro

qwen3.7-max-2026-05-20

qwen3.6-plus-2026-04-02

qwen3.5-plus

qwen3.5-plus-2026-02-15

qwen3-max-2025-09-23

error Error umum