README.md

PromptL

pip install promptl-ai

Requires Python 3.9 or higher.

What is PromptL?

PromptL offers a common, easy-to-use syntax for defining dynamic prompts for LLMs. It is a simple, yet powerful language that allows you to define prompts in a human-readable format, while still being able to leverage the full power of LLMs.

This repository contains the Python bindings for PromptL. Go to the main repository to learn more.

Usage

from promptl_ai import Promptl

promptl = Promptl()

prompt = """
<step>
    First, think step by step about how to answer the user's question.
    <user>
        Taking into account this context: {{context}}
        I have the following question: {{question}}
    </user>
</step>
<step>
    Finally, answer the user's question succinctly yet complete.
</step>
"""

chain = promptl.chains.create(
    prompt=prompt,
    parameters={
        "context": "PromptL is a templating language specifically designed for LLM prompting.",
        "question": "What is PromptL?",
    },
)
conversation = chain.step()
conversation = chain.step("Reasoning...")  # LLM response
conversation = chain.step("Answer...")  # LLM response

print(conversation.messages)

Find more examples.

Development

Requires uv 0.5.10 or higher.

• Install dependencies: uv venv && uv sync --all-extras --all-groups
• Add [dev] dependencies: uv add [--dev] <package>
• Run linter: uv run scripts/lint.py
• Run formatter: uv run scripts/format.py
• Run tests: uv run scripts/test.py
• Build package: uv build
• Publish package: uv publish

License

The PromptL bindings are licensed under the MIT License - read the LICENSE file for details.