Welcome to Switchport

What is Switchport?

Switchport is a platform that helps you manage, test, and optimize LLM prompts in production. Execute prompts, run A/B tests, and track metrics to continuously improve your AI applications.

Key Features

Prompt Management

Store and version your prompts in one place. Update prompts without deploying code.

A/B Testing

Test multiple prompt versions simultaneously with deterministic routing based on subject identification.

Metrics Tracking

Track performance metrics and aggregate results per prompt version automatically.

Multi-Model Support

Support for OpenAI, Anthropic, Google Gemini, and more.

Get Started

Choose your language to begin:

Python SDK

Get started with the Python SDK in 5 minutes.

Go SDK

Get started with the Go SDK in 5 minutes.

TypeScript SDK

Coming soon - TypeScript/JavaScript SDK.

How it Works

Create a prompt config

Define your prompt in the Switchport dashboard with variables and multiple versions.

Execute from your code

Use the SDK to execute prompts with dynamic variables and subject identification for A/B testing.

Track metrics

Record user feedback, conversion rates, or any custom metrics linked to prompt versions.

Analyze results

View aggregated metrics per version in the dashboard to identify the best-performing prompts.

Quick Example

from switchport import Switchport

client = Switchport()

# Execute a prompt
response = client.prompts.execute(
    prompt_key="welcome-message",
    subject={"user_id": "user_123"},
    variables={"name": "Alice"}
)

print(response.text)

# Track user feedback
client.metrics.record(
    metric_key="satisfaction",
    value=4.5,
    subject={"user_id": "user_123"}
)

Resources

GitHub

Explore our open-source SDKs

Dashboard

Access the Switchport platform

Discord

Join our community

​What is Switchport?

​Key Features