OpenAI API + Vol 4: Per-Client Usage Tracking

If you're calling the OpenAI API directly for multiple clients, Vol 4 gives you per-client cost breakdowns with a single line change. Works with the official Python and Node SDKs.

Setup

Step 1

Install the OpenAI SDK

pip install openai (you probably already have this).

Step 2

Set the base URL

Pass base_url when creating the client. That's the only change.

Step 3

Add the client header

Use default_headers to tag every request to a client.

Step 4

Make API calls as usual

Your code doesn't change. Chat completions, embeddings, images — everything routes through Vol 4.

Step 5

See per-client costs

Dashboard shows cost, tokens, model, latency — broken down by client.

Code

from openai import OpenAI

client = OpenAI(
    api_key="sk-your-key",
    base_url="https://proxy.vol4.ai/v1",
    default_headers={
        "X-Vol4-Client": "haufe-group"
    }
)

response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello"}]
)

# That's it. One line changed. Costs tracked per client.

Why this matters

The OpenAI dashboard shows one aggregate number. Vol 4 splits it per client, per model, per day — so you can price accurately and catch margin problems before they cost you.

Start tracking in 5 minutes

Get Early Access