MLflow Access Control (mlflow-oidc-auth)
Authentication and access control for MLflow tracking servers: single sign-on (OIDC, SAML 2.0), SCIM user and group provisioning, service accounts, and per-resource permissions for users, groups and workspaces.
It is an MLflow server plugin, installed as mlflow-oidc-auth and started with --app-name oidc-auth. The package keeps its original name; it has long since grown beyond OIDC.
Disclaimer
This project is not affiliated with, endorsed by, or sponsored by the MLflow Project, Databricks, the Linux Foundation, or LF Projects, LLC. MLflow and related marks are trademarks of their respective owners. Maintained by Kharkevich Engineering Lab.
Features
- Single sign-on for the MLflow UI and API through any OpenID Connect provider (confidential or PKCE public clients) or SAML 2.0 identity provider, with several providers side by side
- Programmatic access: automation authenticates with short-lived workload identities — Kubernetes service-account tokens or IdP client-credentials / workload-identity tokens (JWT bearer); people using the MLflow client from a laptop or notebook use named personal access tokens (basic auth). See Programmatic access
- SCIM 2.0 provisioning of users and groups from your directory
- Permissions (READ, USE, EDIT, MANAGE) on experiments, registered models, prompts, scorers and AI Gateway resources, granted to users, groups or regex patterns, with deny by default
- Workspaces for multi-tenant isolation on a shared MLflow server
- Admin UI for users, groups, service accounts, permissions, workspaces, webhooks and trash
- Operations: server-side sessions, audit log, health probes, Redis-backed permission cache, and secrets from AWS, Azure, HashiCorp Vault or Kubernetes
Documentation
For detailed documentation, please refer to the docs. AI generated documentation is available at DeepWiki.
Quick Start
To get the full version (with entire MLflow and all dependencies), run:
python3 -m venv venv
source venv/bin/activate
python3 -m pip install mlflow-oidc-auth[full]
mlflow server --app-name oidc-auth --host 0.0.0.0 --port 8080
Webhook secret encryption key 🔐
Webhook secrets are stored encrypted in the database using a Fernet key. If you plan to use MLflow webhooks with secrets, set the encryption key in the environment variable MLFLOW_WEBHOOK_SECRET_ENCRYPTION_KEY before creating any webhooks. Generate a key with:
MLFLOW_WEBHOOK_SECRET_ENCRYPTION_KEY=$(python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())")
export MLFLOW_WEBHOOK_SECRET_ENCRYPTION_KEY
Important: keep this key stable across application restarts and replicas. If the key is lost or changed after webhooks are created, previously stored secrets cannot be decrypted and will cause webhook listing to fail until you restore the original key or remove/rotate the affected webhook secrets.
Development
For development quick start, please refer to the Development and Contribution section. Contribution guidelines are available in CONTRIBUTING.md.
License
Apache 2 Licensed. For more information, please see LICENSE.
Based on MLflow basic-auth plugin
https://github.com/mlflow/mlflow/tree/master/mlflow/server/auth
1 maintainer
Metadata
1 maintainer
Release files for mlflow-oidc-auth 9.0.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| mlflow_oidc_auth-9.0.2.tar.gz | 1.5 MB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| mlflow_oidc_auth-9.0.2-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 3.3 MB
Release files / mlflow_oidc_auth-9.0.2.tar.gz
| Download URL | mlflow_oidc_auth-9.0.2.tar.gz |
|---|---|
| Size | 1.5 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3fbb7efd3c34f0172d11a646d2324d8cd4e505566e089c9f216cd50c0f396d8a
|
|
BLAKE2b-256 checksum How to use checksums |
f796983da4f97c3c3e364f2ba707929b460b62aec5f1c3e37832924bea787059
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 1, 2026.
Transparency logRelease files / mlflow_oidc_auth-9.0.2-py3-none-any.whl
| Download URL | mlflow_oidc_auth-9.0.2-py3-none-any.whl |
|---|---|
| Size | 1.8 MB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
a3583ef81e1b94bd7ea7c8926d8255e875c06e49d10cb98e46279bac7210ff57
|
|
BLAKE2b-256 checksum How to use checksums |
cd056775a1a6162e20ea76c1fd0a0eb78e306d29b659f186999d567551decece
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 1, 2026.
Transparency log
