Azure OpenAI

Azure OpenAI lets you keep traffic inside an EU region, bill via an existing Azure contract, and layer Microsoft’s content filters on top of TapPass’s own pipeline.

Requirements

Server-side (your platform team sets these — LiteLLM’s Azure env vars):

AZURE_API_BASE=https://<resource>.openai.azure.com
AZURE_API_KEY=<your-azure-key>
AZURE_API_VERSION=2024-08-01-preview  # or newer

Plus at least one model deployment — for example gpt-4o-mini deployed as prod-chat.

Option A — SDK

from tappass import Agent

agent = Agent("https://tappass.example.com", "tp_...")
response = agent.chat("Hello", model="azure/prod-chat")

The azure/ prefix tells TapPass to route to the Azure deployment of that name.

Option B — OpenAI SDK, zero-code

export OPENAI_BASE_URL=https://tappass.example.com/v1
export OPENAI_API_KEY=tp_...

from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="azure/prod-chat",
    messages=[{"role": "user", "content": "Hello"}],
)

No Azure-specific client or endpoint — TapPass handles the protocol translation.

Data residency

Pick an Azure region close to your users (West Europe, North Europe, Sweden Central). azure/gpt-4o and azure/gpt-4o-mini are on the server’s EU-allowed model list used when TAPPASS_EU_DATA_RESIDENCY=true.

Notes

• Azure’s content filter is additive — if Azure flags a response, TapPass surfaces it as a detection.
• Private networking (Private Endpoints, VNet integration) is supported — TapPass only needs outbound access to the Azure endpoint.