ODD Platform (Open Data Discovery) is an Apache-2.0, Java-based, open-source data discovery and observability platform for data teams that need one place to catalogue, trace, monitor and govern the data moving through their pipelines.

What it is

ODD Platform is the reference implementation of the Open Data Discovery Spec and lives in the data engineering and observability ecosystem. It accumulates scattered metadata into a federated data catalogue, draws end-to-end lineage across data objects and microservices, stores that metadata, and serves it through its own lightweight UI. Ingestion runs as a separate enricher component, so the catalogue acts as a central hub rather than another pipeline stage.

The concrete problem it replaces is the manual, fragmented hunt for answers about data. Rather than chasing owners across chat threads, spreadsheets and wikis to learn where a dashboard's numbers come from, who uses a table, or whether a pipeline has broken, a team points ODD at its sources and reads the answer from the catalogue. The same platform covers governance: tags record how and by whom data is used, so compliance checks and deprecation decisions rest on recorded evidence.

Key capabilities

• Federated data catalogue that accumulates scattered data insights in one place, with an ODD-native lightweight UI for browsing them.
• End-to-end data object lineage, plus a separate end-to-end microservices lineage feature tracking data flow across the whole data landscape.
• Pipeline monitoring with alerting, so broken or stalled pipelines raise warnings instead of staying silent.
• Data Quality Dashboard showing metrics, trends and issues across datasets, with compatibility for Great Expectations and DBT tests and integration with any custom DQ framework.
• Tags and usage transparency for data security and compliance.
• ML experiment result storage that automatically logs experiment parameters.
• Reference Data Management through Lookup Tables, holding key elements such as currency codes, country names and product categories as a single source of truth for pipelines and transformations.

Who uses it and how

• Data platform and data engineering teams that need a shared catalogue across many sources instead of one catalogue per team.
• Governance and compliance workstreams that tag how and by whom data is used, and that deprecate outdated objects only after assessing the risk.
• Data quality owners running Great Expectations or DBT tests who want results in a dashboard rather than separate reports.
• Master data management efforts that centralise lookup tables and wire them into pipelines for enrichment and validation.
• Kubernetes-based organisations that deploy the platform through the project's Helm charts.

Getting started

Run the published container image ghcr.io/opendatadiscovery/odd-platform:latest against an existing PostgreSQL instance by setting SPRING_DATASOURCE_URL, SPRING_DATASOURCE_USERNAME and SPRING_DATASOURCE_PASSWORD, or start a full local stack with docker-compose -f docker/demo.yaml up -d odd-platform-enricher. Helm charts and further example configurations in docker/examples cover other setups, and a hosted demo runs at demo.oddp.io.

How it compares

The facts provided name no comparable products or paid alternatives, so ODD Platform stands alone in this registry as the only data discovery and observability platform listed. Its position rests on being the first open-source platform of its kind, released under Apache-2.0 with no vendor product named as the thing it replaces.

When to use it — and when not to

A self-hoster must supply and operate a PostgreSQL database, whether the platform runs as a container or under Helm, so teams without an existing Postgres instance take on that dependency. Teams that want a hosted, zero-operations catalogue, or that have no metadata sources to connect through adapters, will get little from it. The repository carries 135 open issues and the facts state no release version, so confirm current stability before committing.

export POSTGRES_HOST=172.17.0.1
export POSTGRES_PORT=5432
export POSTGRES_DATABASE=postgres
export POSTGRES_USER=postgres
export POSTGRES_PASSWORD=mysecretpassword

docker run -d \
  --name odd-platform \
  -e SPRING_DATASOURCE_URL=jdbc:postgresql://${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DATABASE} \
  -e SPRING_DATASOURCE_USERNAME=${POSTGRES_USER} \
  -e SPRING_DATASOURCE_PASSWORD=${POSTGRES_PASSWORD} \
  -p 8080:8080 \
  ghcr.io/opendatadiscovery/odd-platform:latest

docker-compose -f docker/demo.yaml up -d odd-platform-enricher