Prompty Python SDK

Python client for interacting with the Prompty API to fetch real-time prompts for AI agents.

Installation

pip install prompty-sdk

Or install from source:

cd python-sdk
pip install -e .

Usage

from prompty import PromptyClient

# Initialize the client
client = PromptyClient(
    base_url='https://your-prompty-instance.com',
    project_id='your-project-id',
    api_key='pk_your_api_key_here'
)

# Get the active prompt for an agent
prompt = client.get_prompt('customer-support-agent')
print(prompt)
# Output: "You are a helpful customer support assistant..."

# Get detailed information about the prompt
details = client.get_prompt_details('customer-support-agent')
print(details['prompt_text'])
print(details['prompt_id'])
print(details['updated_at'])

API Reference

PromptyClient

Initialize a Prompty client.

Parameters:

• base_url (str): The base URL of your Prompty instance
• project_id (str): Your project ID
• api_key (str): Your API key

get_prompt(agent_name)

Get the active prompt text for an agent.

Parameters:

• agent_name (str): The name of the agent

Returns:

• str: The prompt text

Raises:

• ValueError: If the API key is invalid or agent not found
• RuntimeError: If there's a server error

get_prompt_details(agent_name)

Get detailed information about the active prompt.

Parameters:

• agent_name (str): The name of the agent

Returns:

• dict: Dictionary containing:
  • prompt_text (str): The prompt text
  • prompt_id (str): The prompt ID
  • updated_at (str): When the prompt was created/updated
  • agent_name (str): The agent name

Raises:

• ValueError: If the API key is invalid or agent not found
• RuntimeError: If there's a server error

Example: Using with OpenAI

from prompty import PromptyClient
import openai

# Initialize Prompty client
prompty = PromptyClient(
    base_url='https://your-prompty-instance.com',
    project_id='your-project-id',
    api_key='pk_your_api_key_here'
)

# Get the prompt from Prompty
system_prompt = prompty.get_prompt('customer-support-agent')

# Use it with OpenAI
response = openai.ChatCompletion.create(
    model="gpt-4",
    messages=[
        {"role": "system", "content": system_prompt},
        {"role": "user", "content": "How do I reset my password?"}
    ]
)

print(response.choices[0].message.content)

Error Handling

The SDK raises the following exceptions:

• ValueError: When the API key is invalid or the requested agent/prompt is not found
• RuntimeError: When there's a server error or network issue

try:
    prompt = client.get_prompt('my-agent')
except ValueError as e:
    print(f"Invalid request: {e}")
except RuntimeError as e:
    print(f"Server error: {e}")

License

MIT