Skip to content

lakeFS

lakeFS provides Git-like version control and management for data lakes built on object storage, enabling isolated experimentation, data governance, and atomic operations on petabytes of data.

Freemium

stork
I am a painting.
One still painting and one typed sentence.
This painting was never filmed.
A photo, a cartoon, a painting or a 3D character. From 10 cents a minute. Free to try.
Make a photo talkDevelopers: there is an API

What is lakeFS?

lakeFS is an AI-ready data control plane tool developed by Treeverse that enables data and AI teams to manage data with software engineering best practices. It provides Git-like version control and management for data lakes built on object storage, supporting isolated experimentation and atomic operations on petabytes of data. The platform functions as a metadata layer above existing cloud storage services such as AWS S3, Google Cloud Storage, and Azure Blob Storage, or S3-compatible systems, without requiring data format migrations or duplicating data. This abstraction facilitates data lifecycle management, collaboration, and reproducible ML workflows.

Key Features of lakeFS

lakeFS transforms object storage into a Git-like repository, offering a suite of features designed to manage data with software engineering best practices and enhance data lake operations.

  • Git-like version control for data lakes, including branching, committing, merging, and reverting.
  • Management for data lakes built on object storage, supporting AWS S3, Google Cloud Storage, Azure Blob Storage, and S3-compatible systems.
  • Isolated experimentation with zero-copy branching on petabytes of data, eliminating data duplication costs.
  • Data governance capabilities, including data lifecycle management and provenance tracking.
  • Atomic operations for consistent data updates and changes.
  • Support for reproducible ML workflows, ensuring data consistency across model iterations.
  • A comprehensive data lake versioning platform accessible via API.
  • Automated data quality validation through commit and merge workflows.
  • Instant rollback functionality to any previous point in time for disaster recovery and error correction.

Who Should Use lakeFS?

lakeFS is designed for data and AI teams seeking to apply software engineering principles to their data management, particularly within large-scale data lake environments.

  • ML/AI Teams: For ensuring ML reproducibility, accelerating AI delivery, and managing large, multimodal datasets with version control.
  • Data Engineering Teams: For creating isolated development and testing environments, automating data quality validation, and streamlining ETL job development.
  • Organizations with Petabyte-Scale Data Lakes: For applying Git-like version control to massive datasets on object storage without incurring high storage duplication costs.
  • Teams Requiring Data Governance and Disaster Recovery: For establishing data sovereignty, auditability, and the ability to instantly roll back to previous data states.

How to Use lakeFS

Getting started with lakeFS involves deploying it as a metadata layer over existing object storage and then utilizing its Git-like operations via API or client.

  1. Deploy lakeFS as a metadata layer over an existing object storage service (e.g., AWS S3, Google Cloud Storage).
  2. Create a lakeFS repository to manage specific data lake assets or an entire data lake.
  3. Utilize the lakeFS API or client to create instant, zero-copy branches of production data for isolated experimentation or development.
  4. Perform data transformations, ETL jobs, or ML model training within the isolated branch without affecting the main data.
  5. Commit changes to the branch, creating a new, versioned state of the data.
  6. Validate changes and data quality within the branch, then atomically merge validated changes back to the main branch or revert if issues are identified.

lakeFS Pricing & Plans

lakeFS operates on a freemium model, offering both an open-source version for community use and an enterprise version with additional features and support.

  • Open Source: Free, available on GitHub (https://github.com/treeverse/lakeFS) for self-managed deployments.
  • Enterprise: Contact Sales for pricing details, which includes advanced features, dedicated support, and enterprise-grade capabilities.

Pros

  • Offers zero-copy branching for cost-effective isolated experimentation on petabytes of data, significantly reducing storage costs.
  • Applies Git-like operations (branch, commit, merge, revert) directly to data lakes on object storage, integrating software engineering best practices.
  • Supports reproducible ML workflows and automated data quality validation, enhancing reliability and debugging for AI projects.
  • Provides instant rollback capabilities for disaster recovery and efficient correction of erroneous or misformatted data.
  • Maintains an active open-source community and is proven in production environments, indicating stability and support.
  • The acquisition of DVC in November 2025 enhances its data version control capabilities and market position.

Cons

  • Requires deployment as a metadata layer over existing object storage, adding an architectural component to the data stack.
  • The Git-like paradigm for data management may require a learning curve and adaptation for data teams accustomed to traditional data lake operations.
  • Specific pricing for the Enterprise version is not publicly disclosed, requiring direct engagement with sales.
  • Mount capabilities for local access to storage data were 'set to launch' in early June (2026), indicating a potential current limitation for certain workflows.
  • Specific numerical user ratings or detailed reviews from platforms like G2 or Capterra are not readily available in the provided data.

lakeFS vs Competitors

lakeFS positions itself as a control plane for AI-ready data, differentiating through its Git-like branching and versioning applied at petabyte scale directly to object storage.

lakeFS

Freemium

Enjoying this? Get one like it in your inbox each morning.

one email a day · unsubscribe in two clicks · no third-party tracking