OpenPipe LLM

The OpenPipeLLMService extends the standard OpenAI service to integrate with OpenPipe's powerful fine-tuning and monitoring platform. It allows developers to log requests and apply tags for easy dataset collection and model evaluation.

Installation

To use OpenPipe, install the required dependencies:

pip install "piopiy-ai[openpipe]"

Prerequisites

An OpenPipe API key (Get yours here).

Set your keys in your environment:

export OPENPIPE_API_KEY="your_openpipe_key_here"
export OPENAI_API_KEY="your_openai_key_here"

Configuration

`OpenPipeLLMService` Parameters

Parameter	Type	Default	Description
`model`	`str`	`"gpt-4.1"`	The model name (OpenAI or fine-tuned).
`api_key`	`str`	`None`	OpenAI API key (falls back to env).
`openpipe_api_key`	`str`	`None`	OpenPipe API key (falls back to env).
`openpipe_base_url`	`str`	Default	OpenPipe API endpoint URL.
`tags`	`dict`	`None`	Metadata tags for tracking requests.

Usage

Basic Setup with Tagging

import os
from piopiy.services.openpipe.llm import OpenPipeLLMService

llm = OpenPipeLLMService(
    model="gpt-4.1",
    openpipe_api_key=os.getenv("OPENPIPE_API_KEY"),
    tags={
        "environment": "production",
        "app_version": "1.2.0",
        "experiment_id": "voice-v1"
    }
)

Notes

Automatic Logging: All requests made through this service are automatically logged to the OpenPipe dashboard, making it easy to create fine-tuning datasets from production traffic.
OpenAI Compatible: Since it inherits from OpenAILLMService, you can use it as a drop-in replacement for standard GPT models while gaining monitoring benefits.

Installation​

Prerequisites​

Configuration​

OpenPipeLLMService Parameters​

Usage​

Basic Setup with Tagging​

Notes​