Skip to main content

Getting Started with the Hub

1. Install the Hub CLI​

The Guardrails Hub CLI is a part of the Guardrails AI package. You can install it using pip:

pip install guardrails-ai

2. Configure the Hub CLI​

To configure the Hub CLI, you need to set your API key. You can get your API key from the Guardrails Hub website.

guardrails configure

3. Install a Guardrail​

You can download a guardrail from the Hub using the install command. For example, to download the regex_match guardrail, you can run:

gudardrails hub install hub://guardrails/regex_match

4. Use the Guardrail​

You can use the guardrail in your code by creating a Guard object and passing it to your LLM API calls. For example, to use the regex_match guardrail with OpenAI's GPT-3, you can run:

# Import Guard and Validator
from guardrails.hub import RegexMatch
from guardrails import Guard

# Initialize the Guard with 
val = Guard().use(
    RegexMatch(regex="^[A-Z][a-z]*$")
)

guard.parse("Caesar")  # Guardrail Passes
guard.parse("Caesar is a great leader")  # Guardrail Fails

5. Run multiple Guardrails​

You can run multiple guardrails in a single Guard object. For example, here's how to make sure any LLM generated text doesn't contain any toxic language or any mention of your competitors:

First, install the necessary guardrails:

guardrails hub install hub://guardrails/competitor_check
guardrails hub install hub://guardrails/toxic_language
from guardrails import Guard
from guardrails.hub import CompetitorCheck, ToxicLanguage

competitors = ["Apple", "Samsung"]

guard = Guard().use(
    CompetitorCheck(
        competitors=[competitors]
).use(
    ToxicLanguage(
        validation_method=sentence,
        threshold=0.5
    )
)

guard.validate("My favorite phone is BlackBerry.")  # Guardrail Passes