Prompts

Prompt Templates

Basic Template

Create a reusable template with dynamic placeholders:

from SimplerLLM.prompts.prompt_builder import create_prompt_template

template = create_prompt_template("Explain {topic} in {style} terms.")
prompt = template.assign_parms(topic="machine learning", style="simple")
print(prompt)  # "Explain machine learning in simple terms."

Update the template at any time:

template.update_template("Summarize {topic} for a {audience} audience.")
prompt = template.assign_parms(topic="AI safety", audience="technical")

Multiple Values

Generate multiple prompts from one template:

from SimplerLLM.prompts.prompt_builder import create_multi_value_prompts

template = create_multi_value_prompts("Write a {length} {type} about {subject}.")

prompts = template.generate_prompts([
    {"length": "short", "type": "blog post", "subject": "AI"},
    {"length": "long", "type": "article", "subject": "machine learning"},
])

for prompt in prompts:
    print(prompt)

Messages Template

Build validated conversation message lists for chat-based LLMs:

from SimplerLLM.prompts.messages_template import MessagesTemplate

msgs = MessagesTemplate()
msgs.add_user_message("What is Python?")
msgs.add_assistant_message("Python is a programming language.")
msgs.add_user_message("What is it used for?")

messages = msgs.get_messages()
# [
#   {"role": "user", "content": "What is Python?"},
#   {"role": "assistant", "content": "Python is a programming language."},
#   {"role": "user", "content": "What is it used for?"}
# ]

Validate message structure before sending:

is_valid, message = msgs.validate_alternation()
print(is_valid)   # True
print(message)    # "MessageTemplate is valid."

Note: Messages must start and end with a user message, and roles must alternate between user and assistant.

Method	Description
add_user_message(content)	Append a user message
add_assistant_message(content)	Append an assistant message
validate_alternation()	Returns (bool, str) — validity and message
get_messages()	Returns validated message list
get_last_message()	Returns last message or None
prepend_messages(messages_list)	Prepend messages to the start

Prompt Hub

Fetch and manage prompts from the SimplerLLM Prompt Hub API.

Setup

Set your API key in .env:

SIMPLERLLM_API_KEY=your_api_key

Fetch a Prompt

from SimplerLLM.prompts.hub import fetch_prompt_from_hub

prompt = fetch_prompt_from_hub("your_prompt_id")

prompt.set_variables(target_audience="developers", tone="friendly")
formatted = prompt.get_formatted_prompt()
print(formatted)

Parameter	Type	Default	Description
prompt_id	str	—	Unique prompt identifier
api_key	str	None	API key (falls back to SIMPLERLLM_API_KEY env var)

List Prompts

from SimplerLLM.prompts.hub import list_prompts_from_hub

prompts = list_prompts_from_hub()

for p in prompts:
    print(f"{p.name} (v{p.version}) - {p.description}")

Parameter	Type	Default	Description
api_key	str	None	API key (falls back to env var)
include_shared	bool	True	Include shared prompts in results

Fetch Specific Version

from SimplerLLM.prompts.hub import fetch_prompt_version_from_hub

prompt = fetch_prompt_version_from_hub("your_prompt_id", version=2)
print(prompt.template)

Parameter	Type	Default	Description
prompt_id	str	—	Unique prompt identifier
version	int	—	Version number to fetch
api_key	str	None	API key (falls back to env var)

ManagedPrompt

The object returned by fetch_prompt_from_hub():

Property	Type	Description
id	str	Prompt ID
name	str	Prompt name
description	str	Prompt description
template	str	Raw template string
variables	list	List of prompt variables
tags	list	Tags
version	int	Version number

Method	Description
set_variable(name, value)	Set a single variable
set_variables(**kwargs)	Set multiple variables at once
get_formatted_prompt()	Returns template with variables substituted