Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.switchport.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Prerequisites

​
Installation

Install the Switchport SDK using pip: 
pip install switchport
Or install from source: 
git clone https://github.com/switchport-ai/switchport-python.git
cd switchport-python
pip install -e .

​
Get Your API Key

1

Sign in to Switchport

Go to switchport.ai and log in to your account.
2

Navigate to Settings

Go to Settings → API Keys in the dashboard.
3

Copy your API key

Copy your API key (it starts with sp_).

​
Set Your API Key

Set your API key as an environment variable: 
export SWITCHPORT_API_KEY=sp_your_key_here

​
Create Your First Prompt

Before using the SDK, you need to create a prompt in the Switchport dashboard:
1

Create a prompt config

  1. Go to Prompts → New Prompt Config
  2. Name: “Welcome Message”
  3. Key: welcome-message
  4. Click Create
2

Add a version

  1. Click Add Version
  2. Model: Select gpt-5 (or another model)
  3. Prompt: Write a friendly welcome message for {{name}}.
  4. Click Save
3

Publish the version

Click Publish on the version you just created.

​
Configure LLM API Keys

The SDK calls LLMs on your behalf, so you need to configure your LLM API keys:
1

Go to Organization Settings

Navigate to Settings → Organization Settings
2

Add your LLM API keys

Add API keys for the LLM providers you want to use:
  • OpenAI API key (for GPT models)
  • Anthropic API key (for Claude models)
  • Google API key (for Gemini models)
3

Save

Click Save to store your API keys securely.

​
Execute Your First Prompt

Create a file test_switchport.py: 
from switchport import Switchport

# Initialize client (reads API key from environment)
client = Switchport()

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

print("Generated text:")
print(response.text)
print(f"\nModel: {response.model}")
print(f"Version: {response.version_name}")
Run it: 
python test_switchport.py
You should see output like: 
Generated text:
Hello Alice! Welcome to our platform. We're excited to have you here!

Model: gpt-5
Version: v1

​
Record Your First Metric

Now let’s track a metric. First, create a metric definition in the dashboard:
1

Create metric definition

  1. Go to Metrics → New Metric
  2. Key: satisfaction
  3. Name: “User Satisfaction”
  4. Type: float
  5. Click Create
Then record a metric in your code: 
from switchport import Switchport

client = Switchport()

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

# Simulate user feedback (1-5 stars)
user_rating = 4.5

# Record metric with same subject
result = client.metrics.record(
    metric_key="satisfaction",
    value=user_rating,
    subject={"user_id": "user_123"}  # Same context!
)

print(f"Metric recorded! Event ID: {result.metric_event_id}")
Always use the same subject when executing prompts and recording metrics. This ensures metrics are correctly aggregated per prompt version.

​
Next Steps

A/B Testing

Learn how to run A/B tests with multiple prompt versions

API Reference

Explore the full API reference

Examples

See more code examples

Core Concepts

Understand key concepts