OpenAI

Two ways in: point the OpenAI client at the proxy, or call init() and keep the client as it is (auto-instrumentation patches the openai package in Python and Node).

Proxy

Proxy Base URL: https://api.promptguard.co/api/v1Point your LLM SDK’s base_url / baseURL to this endpoint. Your existing provider API key goes in the Authorization header as usual; add your PromptGuard key via X-API-Key.

The proxy serves the Chat Completions route (/chat/completions) and the Responses API route (/responses) for OpenAI models. It sends a model to OpenAI when its id starts with gpt-, o1, o3, o4, o5 or chatgpt-, among the other OpenAI families, and that id is not claimed by a more specific provider first (how the proxy picks a provider).

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.promptguard.co/api/v1",
    api_key=os.environ["OPENAI_API_KEY"],  # forwarded to OpenAI
    default_headers={"X-API-Key": os.environ["PROMPTGUARD_API_KEY"]},
)

response = client.chat.completions.create(
    model="gpt-5-nano",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.promptguard.co/api/v1',
  apiKey: process.env.OPENAI_API_KEY, // forwarded to OpenAI
  defaultHeaders: { 'X-API-Key': process.env.PROMPTGUARD_API_KEY },
});

const response = await client.chat.completions.create({
  model: 'gpt-5-nano',
  messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0].message.content);

curl https://api.promptguard.co/api/v1/chat/completions \
  -H "X-API-Key: $PROMPTGUARD_API_KEY" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5-nano", "messages": [{"role": "user", "content": "Hello!"}]}'

To keep the OpenAI key out of your application, store it as a provider key under openai. PromptGuard then sends the stored key and ignores the Authorization header you send. For a GPT-5 model the proxy renames max_tokens to max_completion_tokens, the only name OpenAI accepts for that family, so older client code keeps working.

Auto-instrumentation

import promptguard
promptguard.init()  # reads PROMPTGUARD_API_KEY

from openai import OpenAI
client = OpenAI()  # talks to OpenAI directly; each call is scanned first

Which OpenAI calls are scanned, and which are not, is listed on the auto-instrumentation page.

Azure OpenAI

Azure-hosted OpenAI deployments use their own namespace and headers: see Azure OpenAI.

Get Started

Core Concepts

Gateway

Integrations

Shadow AI

Guides

Cookbooks

Platform

Going to Production

Resources

Roadmap / Preview

Proxy

Auto-instrumentation

Azure OpenAI

​Proxy

​Auto-instrumentation

​Azure OpenAI

Proxy

Auto-instrumentation

Azure OpenAI