LakeLogic
Loading…

Trusted data products,
built by teams and AI agents.

Most data platforms have tables nobody owns, rules nobody enforces and failures found by the consumer. LakeLogic puts the owner, the rules and the service level in a contract, and enforces it from pull request to production.

Run a sample pipeline

Runs in Google Colab. Nothing to install.

One contract runs the pipeline, checks the data and reports how it held up.

Apache 2.0 open sourceRuns in your environmentDatabricksTechnology PartnerMicrosoftISV PartnerNVIDIAInception Program

Works with the stack you already run

  • Polars
  • DuckDB
  • Apache Spark
  • Databricks
  • Microsoft Fabric
  • Snowflake
  • BigQuery
  • Amazon S3
  • Azure ADLS Gen2
  • GitHub
  • GitLab
  • Azure DevOps
  • Bitbucket
  • Jira
  • Slack
  • Microsoft Teams
  • Power BI
  • Tableau

Why LakeLogic?

Every table gets an owner, a rule and a service level, enforced in the pull request, not in a wiki.

Know what exists and what depends on it.

Discover assets, profile data and draft contracts that capture schemas, ownership and dependencies.

Explore your data estate

From data contract to working pipeline.

Design, test and deliver data products on your platform.

  1. Design

    Draft contracts from your sources. Review schemas, quality rules and transformations.

  2. Test before delivery

    Run against synthetic data in Lakehouse Studio. Inspect accepted rows, failures and their reasons.

  3. Deliver

    Create a ready-to-run project for your platform, with contracts, pipeline code and jobs or workflows, opened as a pull request on your repository.

model:
  fields:
  - name: ip_address
    type: string
    required: false
    pii: true
    classification: ip_address
  # ...
primary_key:
- session_id
# ...
quality:
  row_rules:
  - name: session_id_not_null
    sql: session_id IS NOT NULL
From the RideFlow blueprint

Batch, micro-batch and streaming workloads. Streaming runs on Apache Spark. Engine support

Explore a governed data mesh in action.

One set of data contracts builds domain-owned data products on Databricks, Microsoft Fabric and Snowflake. Open any blueprint to see its pipelines, its quality checks and the rows they reject.

RideFlow architecture: YAML contracts and synthetic test data feed LakeLogic Core, which reads sources such as CSV, JSON, PDF, images and DOCX and builds isolated domain products, shared multi-domain gold for consumers, quarantine for rejected rows and run metadata in your platform (Databricks, Microsoft Fabric, Snowflake, AWS Glue or Google BigQuery), with optional telemetry to the LakeLogic Platform.

  • Microsoft Fabric

    A governed data mesh on OneLake, with a lakehouse per domain. Fabric Data Pipelines run each domain from Bronze to Gold, with quality checks and quarantine from the contracts.

    Explore the blueprintPlatform capabilities

  • Databricks

    A governed data mesh on Unity Catalog, with a catalog per domain. Databricks Workflows run each domain from Bronze to Gold, with quality checks and quarantine from the contracts.

    Explore the blueprintPlatform capabilities

  • Snowflake

    A governed data mesh on Snowflake, with a database per domain. Snowflake Tasks run each domain from Bronze to Gold, with quality checks and quarantine from the contracts.

    Explore the blueprintPlatform capabilities

RideFlow uses synthetic data; review each blueprint’s prerequisites before deployment. LakeLogic Core executes supported contracts with Polars, DuckDB and Apache Spark. Compare engine support

Your production data runs in your environment.

Your contracts, your repository, your approval.

  • Production runs in your environment

    LakeLogic Core processes production data in infrastructure you control and reports operational evidence to the platform.

  • You own the contracts

    Contracts remain version-controlled in your repository.

  • You approve changes

    Material changes pass through review and your delivery workflow.

  • Open execution

    LakeLogic Core is open source under Apache 2.0, so you can read the code that runs on your data.

Build with us.

Bring a live data project. Work directly with the founders to design, test and deliver a governed data product, and help shape the platform.

Become a design partner

Get to know LakeLogic

What is LakeLogic?

LakeLogic is a data contract platform that helps teams and AI agents design, deliver and maintain trusted data products. It brings agreed definitions, quality checks and service expectations into the engineering workflow, from testing changes to monitoring production.

What is a data contract?

A data contract defines the agreement between data producers and consumers: the data’s meaning, ownership, quality and service expectations. LakeLogic’s executable contracts also define sources, transformations and how outputs are stored. LakeLogic Core runs the supported processing and checks in your environment.

Does my production data leave my environment?

LakeLogic Core processes production data in infrastructure you control. Only operational metadata, never data values, is sent to the platform for monitoring. Lakehouse Studio can also use samples you choose to upload for design and testing. Use synthetic or appropriately sanitised data.

Which platforms does it work with?

Explore blueprints for Microsoft Fabric, Databricks and Snowflake. LakeLogic Core executes supported contracts with Polars, DuckDB and Apache Spark. Capabilities vary by platform and engine.

Is LakeLogic open source?

LakeLogic Core, the contract execution engine, is open source under Apache 2.0. The hosted platform adds collaborative design, governance and monitoring.

How do I try it?

Run a sample pipeline in Google Colab with nothing to install locally, or to explore LakeLogic in your own workspace.