---
id: valohai-papi
version: "0.1.3"
license: MIT
license_treatment: permissive
maintenance: active
---

# valohai-papi — Experimental imperative Valohai pipeline API

License: permissive · Maintenance: active · Downloads: 82.8K/mo

## What it is and what it does

Papi is an experimental Python library for declaring Valohai pipelines using imperative syntax rather than declarative configuration. It wraps the Valohai platform's pipeline concepts into a programmatic API, letting you define machine learning workflows directly in Python code. The package depends on valohai-yaml for configuration handling and dataclasses for data structure support.

The library is explicitly marked as experimental and subject to API changes. Documentation is minimal—the primary reference is the papi_examples directory in the repository. The package has been maintained in an active state but has not received updates since mid-2021, suggesting either stability or dormancy depending on adoption.

Use it for:

- Define Valohai ML pipelines programmatically in Python instead of writing YAML configuration files.
- Integrate pipeline declaration into Python-based ML workflows for version control and code review.
- Experiment with imperative pipeline APIs as part of Valohai's technical validation process.
- Build dynamic pipelines where structure is determined at runtime based on Python logic.

## Worth the install?

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

Papi provides an imperative Python API for declaring Valohai machine learning pipelines, allowing you to define pipeline workflows programmatically instead of through configuration files.

Yes, if you are actively using Valohai and prefer imperative Python over YAML configuration. The low install friction and permissive license make it a low-risk experiment. However, be aware that the API is explicitly experimental, documentation is sparse, and the package has not been updated since 2021—suitable for technical exploration or established Valohai users, but not for production systems requiring stability guarantees.

## Install

pip install valohai-papi
uv add valohai-papi
poetry add valohai-papi

## Installing valohai-papi

Before you install: Low install friction with only two runtime dependencies (valohai-yaml and dataclasses). Repository remains active with recent commits, though the package itself has not been updated since 2021-07-08.

License in practice: MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects.

Quickstart:

pip install valohai-papi

from valohai_papi import Pipeline

# See papi_examples directory in repository for usage examples

Requires Python >=3.6.1,<4.0.0; documentation is minimal and examples are in the repository's papi_examples directory rather than in published docs.

Verify before relying:

- Whether the experimental API has stabilized or remains subject to breaking changes since 2021.
- Current state of unimplemented features mentioned in the package description.
- Active user base and production readiness beyond technical validation.

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 82.8K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

valohai pipeline api, imperative pipeline declaration, ml workflow definition python, valohai papi, pipeline as code, machine learning pipeline builder, valohai automation, ml-pipelines, valohai, experimental

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