proxera 0.1.0

Drop-in OpenAI SDK replacement that routes through Proxera

Proxera Python Client

Drop-in replacement for the OpenAI Python SDK that routes all requests through Proxera — giving you logging, cost tracking, rate-limit management, and provider failover with zero code changes.

Install

pip install proxera

Quick Start

import proxera

client = proxera.get_client()  # reads PROXERA_API_KEY from env
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)

That's it. Every call goes through Proxera's proxy and shows up in your dashboard.

Configuration

Parameter	Env Variable	Default
proxera_key	PROXERA_API_KEY	(required)
base_url	PROXERA_BASE_URL	https://api.getproxera.com/v1

Constructor arguments

client = proxera.get_client(
    proxera_key="px-live_...",       # or set PROXERA_API_KEY
    tags=["my-app", "experiment-1"], # X-Proxera-Tags header
    team="research",                 # X-Proxera-Team header
    provider="anthropic",            # X-Proxera-Provider header
)

Async

import proxera

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

Wrapper classes

If you prefer an object that holds its config:

client = proxera.Client(tags=["my-app"])
client.chat.completions.create(model="gpt-4o", messages=[...])

Usage with LangChain

from langchain_openai import ChatOpenAI
import proxera

llm = ChatOpenAI(
    model="gpt-4o",
    openai_api_key="px-live_...",
    openai_api_base="https://api.getproxera.com/v1",
)
llm.invoke("Hello!")

Or reuse a Proxera client:

import proxera
from langchain_openai import ChatOpenAI

client = proxera.get_client()
llm = ChatOpenAI(model="gpt-4o", client=client)

Documentation

Full docs at docs.getproxera.com.

License

MIT