---
id: mleap
version: "0.25.2"
license: unclear
license_treatment: permissive
maintenance: active
---

# mleap — MLeap Python API

License: permissive · Maintenance: active · Downloads: 2.3M/mo

## What it is and what it does

MLeap is a Python library that provides serialization and deserialization of machine learning pipelines to a portable Bundle.ML format. It integrates with scikit-learn and enables pipelines built with numpy, scipy, pandas, and scikit-learn to be exported to a common interchange format. The library maintains mathematical parity across implementations, allowing models to produce consistent results when moved between different execution environments.

The package depends on numpy, six, scipy, pandas, and scikit-learn, with optional gensim integration for Word2Vec serialization. It is actively maintained, supports Python 3.10 through 3.13, and carries no known security vulnerabilities. Scikit-Learn deserialization is noted as in development.

Use it for:

- Export a fitted scikit-learn pipeline to Bundle.ML for deployment in a lightweight runtime environment.
- Serialize a gensim Word2Vec model to Bundle.ML for consistent text vectorization across environments.
- Share trained ML pipelines between teams via a standardized interchange format.
- Build reproducible ML workflows where serialized models execute consistently across different contexts.
- Integrate scikit-learn models with systems that consume Bundle.ML format.

## Worth the install?

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

Serializes and deserializes machine learning pipelines to a portable Bundle.ML format for cross-platform deployment and execution.

Yes. MLeap is actively maintained, carries no vulnerabilities, and solves a real portability problem for machine learning pipelines. The low install friction and permissive license make it a straightforward addition. Install if you need to serialize scikit-learn pipelines or integrate with Bundle.ML-based systems; skip if you have no need for model interchange or deployment outside your current framework.

## Install

pip install mleap
uv add mleap
poetry add mleap

## Installing mleap

Before you install: Low friction: pure Python wheel with five standard dependencies (numpy, six, scipy, pandas, scikit-learn). Actively maintained with a release 24 days old and no known vulnerabilities.

License in practice: Permissive Apache license; no restrictions on commercial or proprietary use.

Quickstart:

pip install mleap

from mleap.sklearn.pipeline import to_bundle
import pandas as pd

df = pd.DataFrame([[0.1, 0.2]], columns=['f1', 'f2'])
to_bundle(fitted_pipeline, 'jar:file:/tmp/model.zip', df)

Requires Python 3.10 or later.

Verify before relying:

- Whether deserialization from Bundle.ML to scikit-learn is fully implemented (description indicates it is 'coming soon').
- Whether TensorFlow integration mentioned as 'coming soon' is now available in version 0.25.2.
- Performance characteristics and bundle file size overhead compared to native formats.
- Exact API surface and whether all scikit-learn transformer types are supported.

## Package facts

- License: not declared (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 2.3M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

ml pipeline serialization, model export format, scikit-learn bundle, cross-platform ml deployment, ml model portability, pipeline interchange format, ml interoperability, ml-serialization, model-deployment, pipeline-interchange

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