Quickstart

Create an account

First, navigate to https://lilypad.mirascope.com and create an account.

Environment Variables

Navigate to Settings -> Organization and:

1. Create a new project.
2. Generate an API key for that project.

The API key will only be shown once, so make sure to copy and save it.

LILYPAD_PROJECT_ID="YOUR_PROJECT_ID"
LILYPAD_API_KEY="YOUR_API_KEY"

LILYPAD_PROJECT_ID="YOUR_PROJECT_ID"
LILYPAD_API_KEY="YOUR_API_KEY"

If using a .env file, remember to use something like load_dotenv() so they are properly loaded.

LLM API Key

You'll need to set the API key for the provider you're using in your environment.

We recommend creating one in Google AI Studio if you don't have one yet.

They have a very generous free tier.

Installation

Install the lilypad-sdk with any additional dependencies you may need:

# For spans / tracing
uv add lilypad-sdk

# For Google Gemini/Vertex support
uv add "lilypad-sdk[google]"

# For multiple providers
uv add "lilypad-sdk[google,openai,anthropic]"

# For spans / tracing
uv add lilypad-sdk

# For Google Gemini/Vertex support
uv add "lilypad-sdk[google]"

# For multiple providers
uv add "lilypad-sdk[google,openai,anthropic]"

Available provider extras:

• google - Google Gemini/Vertex models (genai SDK)
• openai - OpenAI models
• anthropic - Anthropic models
• bedrock - AWS Bedrock models
• azure - Azure AI models
• mistral - Mistral models
• outlines - Outlines framework

Automatically Trace & Version

Run your first automatically traced and versioned function:

from google.genai import Client
import lilypad

lilypad.configure(  
    project_id=os.environ["LILYPAD_PROJECT_ID"],
    api_key=os.environ["LILYPAD_API_KEY"],
    auto_llm=True,
)
client = Client()

@lilypad.trace(versioning="automatic")  
def answer_question(question: str) -> str | None:
    response = client.models.generate_content(
        model="gemini-2.0-flash-001",
        contents=f"Answer this question: {question}",
    )
    return response.text
    
response = answer_question("What is the capital of France?")
print(response)
# > The capital of France is Paris.

from google.genai import Client
import lilypad

lilypad.configure(  
    project_id=os.environ["LILYPAD_PROJECT_ID"],
    api_key=os.environ["LILYPAD_API_KEY"],
    auto_llm=True,
)
client = Client()

@lilypad.trace(versioning="automatic")  
def answer_question(question: str) -> str | None:
    response = client.models.generate_content(
        model="gemini-2.0-flash-001",
        contents=f"Answer this question: {question}",
    )
    return response.text
    
response = answer_question("What is the capital of France?")
print(response)
# > The capital of France is Paris.

Follow the link output in your terminal to view the captured version and corresponding trace.

See the observability section to learn more.

Look At Your Data!

Go build and run some LLM functions and inspect their results using Lilypad.

While you're at it, try annotating some data too.