gpubox.ai

Quickstart

Your first call in 60 seconds.

GPUBox is OpenAI-compatible. Any OpenAI SDK works without modification — just point it at https://api.gpubox.ai/v1 and use a GPUBox API key.

  1. Step 1

    Create an API key

    Sign up for GPUBox and create a key from your dashboard. Keys start with gpb_.

    Sign up

  2. Step 2

    Make your first call

    Pick your language. The API surface is identical to OpenAI's — only the base_url changes.

    Python — pip install openai
    from openai import OpenAI

client = OpenAI(
    api_key="gpb_...",
    base_url="https://api.gpubox.ai/v1",
)

response = client.chat.completions.create(
    model="qwen2.5-32b-instruct",
    messages=[{"role": "user", "content": "Hello, GPUBox!"}],
)
print(response.choices[0].message.content)
    Node.js — npm install openai
    import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "gpb_...",
  baseURL: "https://api.gpubox.ai/v1",
});

const response = await client.chat.completions.create({
  model: "qwen2.5-32b-instruct",
  messages: [{ role: "user", content: "Hello, GPUBox!" }],
});

console.log(response.choices[0].message.content);
    curl
    curl https://api.gpubox.ai/v1/chat/completions \
  -H "Authorization: Bearer gpb_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen2.5-32b-instruct",
    "messages": [{"role": "user", "content": "Hello, GPUBox!"}]
  }'

  3. Step 3

    Stream tokens as they arrive

    For chat-like UX, set stream=True. GPUBox emits standard OpenAI SSE chunks.

    Python — streaming
    for chunk in client.chat.completions.create(
    model="qwen2.5-32b-instruct",
    messages=[{"role": "user", "content": "Stream a haiku."}],
    stream=True,
):
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

  4. Step 4

    Send an image (vision)

    Pick the vision model qwen2.5-vl-7b-instruct and add an image_url content part to the message. Same /v1/chat/completions endpoint. The url accepts a base64 data-URI or a plain https image URL.

    Python — vision
    # Vision: send an image alongside your text prompt.
# Use the multimodal model and an image_url content part. The url
# can be a base64 data-URI (shown) or a plain https image URL.
response = client.chat.completions.create(
    model="qwen2.5-vl-7b-instruct",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What's in this screenshot?"},
                {
                    "type": "image_url",
                    "image_url": {"url": "data:image/png;base64,<...>"},
                },
            ],
        }
    ],
)
print(response.choices[0].message.content)

  5. Step 5

    Transcribe audio

    Whisper-large-v3-turbo runs on the same key. Multipart upload, OpenAI-shape response.

    Python — transcription
    with open("audio.wav", "rb") as f:
    transcript = client.audio.transcriptions.create(
        model="whisper-large-v3-turbo",
        file=f,
    )
print(transcript.text)

What's next