---
id: mlserver-mlflow
version: "1.7.1"
license: Apache-2.0
license_treatment: permissive
maintenance: aging
---

# mlserver-mlflow — MLflow runtime for MLServer

License: permissive · Maintenance: aging · Downloads: 80.3K/mo

## What it is and what it does

mlserver-mlflow is a specialized runtime adapter that bridges MLflow and MLServer, allowing MLflow-serialized models to be served as REST/gRPC inference endpoints. It extends MLServer's runtime system with MLflow-specific model loading and inference logic, and introduces a dict content type for handling MLflow's native dictionary-of-tensors input format.

The package is narrow in scope: it does not train models or manage MLflow tracking—it only handles model serving. It is most useful in deployment pipelines where MLflow is the model registry and MLServer is the inference platform. The aging maintenance status (434 days since last release) suggests the package is stable but not actively developed; it will work for its intended use but may lag behind new MLflow or MLServer features.

Use it for:

- Serve MLflow-registered models as production REST endpoints without writing custom inference code.
- Integrate MLflow model artifacts into a MLServer-based inference platform for multi-model serving.
- Deploy dictionary-of-tensors models from MLflow that require the dict content type for correct deserialization.

## Worth the install?

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

Provides a MLServer runtime that loads and serves MLflow models, enabling MLflow-serialized models to run as inference endpoints with support for dictionary-based tensor inputs.

Yes, if you are already using MLServer and need to serve MLflow models. The low install friction and permissive license make it a straightforward addition. However, the aging maintenance status (434 days since release) means you should verify that the package's feature set matches your MLflow and MLServer versions before committing to production use.

## Install

pip install mlserver-mlflow
uv add mlserver-mlflow
poetry add mlserver-mlflow

## Installing mlserver-mlflow

Before you install: Low install friction with a pure Python wheel. Maintenance status is aging—last release was 434 days ago—so expect slower response to issues, though the package remains functional for its narrow purpose.

License in practice: Apache-2.0 permissive license allows commercial and private use with minimal restrictions; suitable for most projects that can comply with attribution and derivative-work disclosure.

Quickstart:

pip install mlserver mlserver-mlflow

from mlserver import MLServer
from mlserver_mlflow import MLflowRuntime

# MLServer will auto-load MLflow models when configured with this runtime

Requires mlserver and mlflow to be installed; Python 3.9–3.12 only (capped below Python 3.13).

Verify before relying:

- Whether the dict content type stacking with other content types (e.g., np) is production-ready or still experimental.
- Current state of the upstream repository and whether maintenance is expected to resume.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: capped_below_current
- Install friction: low
- Maintenance: aging
- Downloads: 80.3K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

mlflow model serving, mlserver runtime mlflow, deploy mlflow models, mlflow inference endpoint, mlflow model deployment, model-serving, mlflow-integration

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