OpenDataDiscovery Platform

OpenDataDiscovery Platform (ODD) is an open-source catalog for data assets and machine-learning metadata. It is best suited to teams that want to connect discovery across data engineering and data science. Its reach across search, lineage, quality, and experiments is a strong reason to consider it, provided your team can manage deployment and authentication.

Overview

ODD brings catalog search together with end-to-end data-object and microservices lineage, imported data-quality results, and machine-learning experiment metadata. Search filters include data source, owner, and tag; users can also inspect ownership and links to related sources. This is useful when finding an asset is only the first step and teams need to understand its dependencies and context.

An open metadata collection standard connects catalogs, tools, and feature stores. Lookup Tables let teams maintain reference data for governance, while owners of affected entities receive alerts after the first ETL execution. The alert timing is useful for follow-up after that run, though it does not describe a broader alert schedule.

Key features

Lineage, quality, and machine learning

Pipeline and report lineage, impact analysis, and lineage across data objects and microservices give teams several ways to trace relationships. ODD can import test results and show dataset quality statuses, so it can make existing quality checks easier to discover; it does not replace the process that produces those results. For data science and engineering teams, logging experiments and cataloging their metadata and lineage can bring ML work into the same discovery process as data assets.

Integrations and deployment

Integrations include MongoDB, ClickHouse, dbt, Snowflake, Tableau, BigQuery, Kafka, and PostgreSQL. A user interface and REST API expose collected metadata. Teams can deploy with local Docker Compose, Kubernetes Helm charts, or AWS Marketplace QuickLaunch, giving different hosting routes while leaving deployment operations with the team.

LDAP authentication is supported, and ODD defines USER and ADMIN roles, with ADMIN holding all permissions by default. Its documentation cautions against using simple login-form authentication in production, so production teams should plan an appropriate authentication setup. Attachments are capped at 20 MB; larger or remotely stored files can be linked instead.

Pricing

The Open-source platform plan is free. ODD is described as free and open source, with public development. The free plan has no stated seat or usage quota, and a free trial is also offered. There is no paid plan or renewal term described, so this is a natural fit for teams prepared to run and maintain the platform themselves.

Platforms

ODD supports API, Linux, self-hosted, and web use. Its Docker Compose, Kubernetes, and AWS Marketplace deployment options suit teams choosing their own operating environment.

Who it's for

ODD targets data teams of any size, including enterprises, small and midsize companies, and data science teams. It is particularly well matched to organizations that want data engineers and data scientists to share discovery of assets, lineage, quality status, and experiments. Teams that need a managed service or rely on attaching large files should weigh the operational work and 20 MB attachment cap.

Pros and cons

• Pros: Search, lineage, imported quality results, and ML experiment metadata cover more of the discovery process than a dataset index alone.
• Pros: Integrations, an API, and several self-hosted deployment routes give teams options for connecting and running the platform.
• Pros: Open-source development and a free plan make it possible to adopt without a software subscription.
• Cons: Production authentication needs planning because simple login-form authentication is discouraged for production use.
• Cons: Attachments stop at 20 MB, requiring links for larger files.
• Cons: Teams take on deployment and platform operations with the available deployment models.

Alternatives

For a broader comparison, see Data Lineage Tools. Choose Atlas if you want an open-core product with a freemium plan and platforms that include macOS and Windows. Grai is another free, open-source, self-hosted option for teams that want to retain control of their data and hosting environment.

Coalesce Catalog may suit a solo developer looking for a free plan with one user, one project, one environment, and 2,000 capped monthly actions; its paid Starter plan is $150 a month. Choose Databricks Notebooks when a web-based notebook and a free serverless workspace with limited compute better fit the need. Apache Atlas is a free alternative for teams seeking open metadata governance software.

Collate may fit teams considering an open-source core with a separately licensed UI and connectors. Amazon DataZone is an option for teams comfortable with pay-as-you-go charges tied to requests, metadata storage, compute, and AI recommendation tokens. Egeria is another free option.

Verdict

Choose ODD if your data team wants a free, open-source discovery platform that brings catalog search, lineage, quality results, and ML experiment metadata together. Its strongest case is shared visibility across data engineering and data science. Look elsewhere if your team cannot take on deployment and production authentication work, or if metadata attachments larger than 20 MB are central to your workflow.