---
id: poml
version: "0.0.8"
license: The MIT License (MIT) Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software… (full text in the JSON record)
license_treatment: unclear
maintenance: aging
---

# poml — Prompt Orchestration Markup Language

License: unclear · Maintenance: aging · Downloads: 74.5K/mo

## What it is and what it does

POML is a markup language designed to bring structure and reusability to prompt engineering for LLMs. It uses an HTML-like syntax with semantic components (role, task, output-format, etc.) to organize prompt logic, embed external data (text, spreadsheets, images), and apply CSS-like styling to control presentation without changing core content. The Python SDK integrates POML into your application workflow, rendering templates with variables, loops, and conditionals to generate data-driven prompts dynamically.

The package targets developers who write complex, multi-part prompts and want to separate content from formatting, reduce duplication, and manage prompt variations systematically. It requires Python 3.9 or later and depends on pydantic for validation and nodejs-wheel for the underlying rendering engine. The project is maintained by Microsoft but is still in early releases (0.0.8), so the API and feature set may evolve.

Use it for:

- Build reusable prompt templates with variables and conditionals to generate different prompts from a single POML file.
- Embed images, CSV data, or text files directly into prompts and control their formatting via styling rules.
- Separate prompt logic from presentation so you can test different verbosity or syntax styles without rewriting core prompts.
- Organize multi-step reasoning prompts (role, context, task, output format) in a structured, maintainable markup file.
- Integrate POML rendering into a Python application to dynamically generate and send prompts to an LLM API.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

POML is a markup language and Python SDK for structuring, templating, and styling prompts for large language models, with support for embedding external data and managing presentation variations.

Yes, if you regularly write complex, multi-part prompts and want a structured, reusable approach to prompt engineering. The MIT license is permissive, and the package has no known vulnerabilities. However, be aware that it is in early releases (0.0.8) with aging maintenance (354 days since last release), so expect the API to evolve and support to be slower. The nodejs-wheel dependency adds medium install friction. Start with a small pilot to verify it fits your workflow before committing to it in production.

## Install

pip install poml
uv add poml
poetry add poml

## Installing poml

Before you install: Medium install friction due to compiled wheels across multiple platforms (macOS, Linux, Windows). The package depends on nodejs-wheel and pydantic. Maintenance status is aging—last release was 354 days ago—so expect slower response to issues.

License in practice: MIT License permits unrestricted use, modification, and distribution with minimal restrictions. No commercial or proprietary concerns.

Quickstart:

pip install poml

from poml import PomlRenderer

renderer = PomlRenderer()
with open('example.poml', 'r') as f:
    prompt_text = renderer.render(f.read())

Requires Python 3.9 or later. The nodejs-wheel dependency must be installed successfully, which may require platform-specific wheels (macOS, Linux, Windows).

Verify before relying:

- Actual stability and maturity of the 0.0.8 release—early version numbers suggest the API may still be in flux.
- Whether the nodejs-wheel dependency adds significant overhead or complexity to the installation.
- Real-world performance and compatibility with current LLM APIs (OpenAI, Azure, Google) mentioned in the docs.

## Package facts

- License: The MIT License (MIT) Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software… (full text in the JSON record) (unclear)
- Python support: supports_current
- Install friction: medium
- Maintenance: aging
- Downloads: 74.5K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

prompt engineering markup language, llm prompt templating, structured prompt management, prompt styling and formatting, data-driven prompt generation, llm-prompting, template-engine, markup-language

[View on SkillFed](https://skillfed.io/packages/poml) · [View on PyPI](https://pypi.org/project/poml/)