ContextQMD
Back to Opik

Observability for Vercel AI Gateway with Opik

apps/opik-documentation/documentation/fern/docs/tracing/integrations/vercel-ai-gateway.mdx

2.0.22-6605-merge-20655.5 KB
Original Source

Vercel AI Gateway provides a unified interface to access multiple AI providers with edge-optimized performance, built-in caching, and comprehensive analytics. It's designed to work seamlessly with Vercel's edge infrastructure.

Gateway Overview

Vercel AI Gateway provides enterprise-grade features for managing AI API access, including:

  • Unified API: Access hundreds of models across providers via Vercel's AI Gateway with minimal code changes
  • Transparent Pricing: No markup on tokens with "Bring Your Own Key" support
  • Automatic Failover: High reliability with requests routed to alternate providers if primary is unavailable
  • Low Latency: Sub-20ms routing latency through Vercel's global edge network
  • Intelligent Caching: Reduce costs and improve response times with smart caching

Account Setup

Comet provides a hosted version of the Opik platform. Simply create an account and grab your API Key.

You can also run the Opik platform locally, see the installation guide for more information.

Getting Started

Installation

First, ensure you have both opik and openai packages installed:

bash
pip install opik openai

Configuring Opik

Configure the Opik Python SDK for your deployment type. See the Python SDK Configuration guide for detailed instructions on:

  • CLI configuration: opik configure
  • Code configuration: opik.configure()
  • Self-hosted vs Cloud vs Enterprise setup
  • Configuration files and environment variables

Configuring Vercel AI Gateway

You'll need to set up the Vercel AI Gateway in your Vercel project. Follow the Vercel AI Gateway setup guide to configure your gateway.

Set your API keys as environment variables:

bash
export VERCEL_AI_GATEWAY_URL="YOUR_VERCEL_AI_GATEWAY_URL"
export OPENAI_API_KEY="YOUR_OPENAI_API_KEY"

Or set them programmatically:

python
import os
import getpass

if "VERCEL_AI_GATEWAY_URL" not in os.environ:
    os.environ["VERCEL_AI_GATEWAY_URL"] = input("Enter your Vercel AI Gateway URL: ")

if "OPENAI_API_KEY" not in os.environ:
    os.environ["OPENAI_API_KEY"] = getpass.getpass("Enter your OpenAI API key: ")

Logging LLM Calls

Since Vercel AI Gateway provides an OpenAI-compatible API, we can use the Opik OpenAI SDK wrapper to automatically log Vercel AI Gateway calls as generations in Opik.

Simple LLM Call

python
import os
from opik.integrations.openai import track_openai
from openai import OpenAI

# Create an OpenAI client with Vercel AI Gateway's base URL
client = OpenAI(
    api_key=os.environ["OPENAI_API_KEY"],
    base_url=os.environ["VERCEL_AI_GATEWAY_URL"]
)

# Wrap the client with Opik tracking
client = track_openai(client, project_name="vercel-ai-gateway-demo")

# Make a chat completion request
response = client.chat.completions.create(
    model="gpt-3.5-turbo",
    messages=[
        {"role": "system", "content": "You are a knowledgeable AI assistant."},
        {"role": "user", "content": "What is the largest city in France?"}
    ]
)

# Print the assistant's reply
print(response.choices[0].message.content)

Advanced Usage

Using with the @track decorator

If you have multiple steps in your LLM pipeline, you can use the @track decorator to log the traces for each step. If Vercel AI Gateway is called within one of these steps, the LLM call will be associated with that corresponding step:

python
import os
from opik import track
from opik.integrations.openai import track_openai
from openai import OpenAI

# Create and wrap the OpenAI client with Vercel AI Gateway's base URL
client = OpenAI(
    api_key=os.environ["OPENAI_API_KEY"],
    base_url=os.environ["VERCEL_AI_GATEWAY_URL"]
)
client = track_openai(client)

@track
def generate_response(prompt: str):
    response = client.chat.completions.create(
        model="gpt-3.5-turbo",
        messages=[
            {"role": "system", "content": "You are a knowledgeable AI assistant."},
            {"role": "user", "content": prompt}
        ]
    )
    return response.choices[0].message.content

@track
def refine_response(initial_response: str):
    response = client.chat.completions.create(
        model="gpt-4",
        messages=[
            {"role": "system", "content": "You enhance and polish text responses."},
            {"role": "user", "content": f"Please improve this response: {initial_response}"}
        ]
    )
    return response.choices[0].message.content

@track(project_name="vercel-ai-gateway-demo")
def generate_and_refine(prompt: str):
    # First LLM call: Generate initial response
    initial = generate_response(prompt)
    
    # Second LLM call: Refine the response
    refined = refine_response(initial)
    
    return refined

# Example usage
result = generate_and_refine("Explain quantum computing in simple terms.")

The trace will show nested LLM calls with hierarchical spans.

Further Improvements

If you have suggestions for improving the Vercel AI Gateway integration, please let us know by opening an issue on GitHub.