// Trust & Security Report

    MLflow logo

    MLflow

    Open-source, self-hosted AI/ML engineering platform (Apache 2.0): experiment tracking, LLM/agent tracing & observability, evaluation, prompt management, model registry, and an AI Gateway. Originally created by Databricks, donated to the Linux Foundation in 2020. No MLflow-operated hosted SaaS product exists for the core open-source project; third parties (e.g. Databricks, Nebius) separately sell managed-hosting offerings built on top of it.

    Certifications held

    0

    Maturity

    Unknown

    Trains on your data

    No

    Trust center

    No

    // Certification ledger

    Each held certification is backed by a verbatim quote from the vendor's own trust or security page. “Not confirmed” means we could not verify it publicly, not that the vendor lacks it.

    > Show 9 unconfirmed / not-held certifications
    SOC 2 (Type 1/2)
    NOT CONFIRMED

    No public evidence. mlflow.org has no trust center, security page, or compliance page; the project's GitHub security policy covers only vulnerability disclosure and does not mention SOC 2.

    source: mlflow.org
    ISO 27001
    NOT CONFIRMED

    No public evidence of an ISO 27001 certification for the MLflow project.

    source: mlflow.org
    ISO 27017 / 27018 / 27701
    NOT CONFIRMED

    No public evidence.

    source: mlflow.org
    ISO/IEC 42001 (AI management systems)
    NOT CONFIRMED

    No public evidence of an AI-governance certification.

    source: mlflow.org
    CSA STAR
    NOT CONFIRMED

    No public evidence.

    source: mlflow.org
    GDPR (posture)
    NOT CONFIRMED

    No MLflow-specific GDPR posture page. The only located privacy document, LF Projects' general privacy policy, covers website/cookie data collection on lfprojects.org and Linux Foundation project sites, not processing of user model/experiment data by the self-hosted software: "describes policies and procedures about the collection, use, disclosure and sharing of personal information when you use their websites... and participate in or use their Project sites."

    source: lfprojects.org
    HIPAA
    NOT CONFIRMED

    No public evidence of a HIPAA compliance statement or BAA offering; no MLflow-operated hosted service exists to sign a BAA for.

    source: mlflow.org
    PCI DSS
    NOT CONFIRMED

    No public evidence. Not applicable to a self-hosted OSS tracking/observability tool with no payment processing.

    source: mlflow.org
    FedRAMP
    NOT CONFIRMED

    No public evidence.

    source: mlflow.org

    // Privacy & AI training

    Trains on customer data

    No

    Data processing agreement

    Not offered

    Data region

    User-controlled; self-hosted deployment means data residency is determined entirely by where the customer runs their own MLflow server/infrastructure. No MLflow-operated data centers exist for the core open-source product.

    MLflow is self-hosted infrastructure software, not a hosted AI/ML service: users run MLflow on their own servers/cloud accounts, and their experiment data, traces, prompts, and models never transit an MLflow-operated backend. There is no MLflow entity that could train on customer data because there is no MLflow-run product tier that ingests it. This differs from third-party managed offerings (e.g. Databricks Managed MLflow), which are separate vendor products with their own data-handling terms.

    // Security controls

    Encryption in transit / at rest

    Not applicable at the vendor level; self-hosted deployment means encryption is configured and controlled entirely by the customer's own infrastructure (e.g. TLS termination, storage encryption on their chosen backend).

    mlflow.org

    Vulnerability disclosure

    Managed via GitHub's coordinated/private vulnerability reporting process on the mlflow/mlflow repository; the project explicitly does not accept reports via bug-bounty platforms like Huntr.

    github.com

    Project governance

    Hosted under the Linux Foundation as "MLflow Project, a Series of LF Projects, LLC" - a vendor-neutral open governance model, Apache 2.0 licensed.

    mlflow.org

    Known CVE history

    MLflow has had publicly disclosed CVEs in the past (e.g. hard-coded default credentials, auth bypass, path traversal in artifact handling) that were patched via the standard OSS release process; this is normal for widely-used self-hosted OSS and is not a certification gap, but customers self-hosting MLflow are responsible for patching and hardening their own deployment.

    github.com

    // Products & data scope

    MLflow (Open Source Core)Self-hosted ML/LLM/agent tracking, tracing, evaluation, prompt management, and model registry

    Data it handles: Fully self-hosted; customer controls all infrastructure, storage, and data. No data sent to an MLflow-operated backend.

    Free, Apache 2.0 licensed.

    Databricks Managed MLflowThird-party managed/hosted version of MLflow

    Data it handles: Customer data processed within Databricks' hosted platform, subject to Databricks' own compliance program (not MLflow project's).

    Separate commercial vendor (Databricks, Inc.). SOC 2, ISO 27001, and HIPAA certifications for this offering belong to Databricks, not to the MLflow open-source project.

    Nebius Managed Service for MLflowThird-party managed/hosted version of MLflow

    Data it handles: Customer data processed within Nebius' cloud, subject to Nebius' own terms/compliance program.

    Separate commercial vendor (Nebius). Same host-vs-vendor caveat as Databricks Managed MLflow.

    // What to watch

    • MLflow is an open-source, self-hosted project (Apache 2.0, governed by LF Projects, LLC / Linux Foundation), not a commercial SaaS vendor. There is no trust center, SOC2/ISO/HIPAA certification, or DPA for the core open-source software because there is no MLflow-operated service that processes customer data.
    • Managed/hosted MLflow offerings exist from third parties (notably Databricks Managed MLflow and Nebius). Compliance certifications those hosts hold belong to the host, not to the MLflow project.
    • The only privacy document found (LF Projects' privacy policy) governs website/cookie data on lfprojects.org and Linux Foundation project sites in general, not the software's handling of user experiment/model data - it should not be read as a data-processing privacy policy for the MLflow product itself.
    • The maturity taxonomy (enterprise/growth/startup) is a poor fit for a widely-adopted (30M+ downloads/month, Linux Foundation-governed) open-source project with no commercial vendor entity, so maturity is shown as unknown rather than forced into a SaaS-vendor category.

    // At a glance

    Pricing model

    Free and open source (Apache 2.0) for the core self-hosted project; paid managed-hosting alternatives are sold separately by third parties (Databricks, Nebius, and others)

    Self-hostable

    Yes

    // How we verified this

    Every certification marked HELD is confirmed against a verbatim quote on MLflow Project (a Series of LF Projects, LLC / Linux Foundation)'s own trust, security, or privacy pages. We reject certifications claimed only on third-party aggregators, on a cloud host's behalf, or by a similarly named company.

    Last verified 2026-07-08. Compliance changes over time. Always confirm directly with the vendor before relying on any certification for a purchasing or compliance decision.

    lfprojects.org

    > Browse all vendor trust reports