Amundsen
Amundsen is an open-source data discovery and metadata platform created by Lyft and now hosted by the LF AI & Data Foundation as an incubation-stage project.
Profile
Open-source metadata search and discovery platform for finding and cataloging data assets across organizations.
Amundsen is an open-source data discovery and metadata platform created by Lyft and now hosted by the LF AI & Data Foundation as an incubation-stage project. The project was started in 2019 by engineers including Tao Feng and Mark Grover to solve a core problem: enabling data analysts, scientists, and engineers to rapidly find and understand data within large organizations. The platform indexes tables, dashboards, and other data resources, then powers discovery through a PageRank-style search algorithm that ranks results by usage patterns—higher-queried tables surface first.
Amundsen's appeal lies in its architectural simplicity. It runs on two main components: a metadata service backed by Neo4j or Apache Atlas for storage, and an Elasticsearch index for search. This minimal stack means faster deployment than competitors and lower operational burden. The platform integrates with major data warehouses including Snowflake, BigQuery, Redshift, and Hive, and is used by more than 30 organizations including ING, Square, Workday, Snap, Instacart, Asana, and Brex.
The project declined materially after 2023, when Stemma—a commercial managed version inspired by Amundsen and launched by co-creator Mark Grover—was acquired by Teradata. Stemma's maintainers moved to Teradata, leaving Amundsen in a dormant state. The most recent software release was August 2024, and as of early 2026, the project has not graduated from LF AI & Data incubation, has no published roadmap, and competes in a crowded market against actively developed alternatives like DataHub and OpenMetadata. For existing users, Amundsen remains stable and capable. For new deployments, community consensus points toward more actively maintained tools.
Who buys this
- Data teams at mid-to-large tech companies running mature data warehouses
- Organizations with Snowflake, BigQuery, or Redshift who need centralized metadata discovery
- Enterprises needing lightweight, self-hosted data catalogs without complex governance
- Data mesh practitioners requiring simple ownership and usage tracking for datasets
Publicly disclosed clients
- Lyft
- Square
- Instacart
- Snap
- ING
- Workday
- Asana
- Brex
Strengths and what to watch
Strengths
- Simple, lightweight architecture with fewer dependencies than DataHub, enabling faster self-hosted deployment
- PageRank-style search algorithm surfaces frequently used data first, improving discovery for large organizations
- Wide integration support with major data warehouses (Snowflake, BigQuery, Redshift, Hive, PostgreSQL)
Watch for
- No active development since August 2024; project maintainers moved to Teradata after Stemma acquisition in 2023
- Stale documentation and no published roadmap; has not graduated from LF AI & Data incubation stage
- Missing advanced features like native column-level lineage and sophisticated data governance, pushing new users toward DataHub or OpenMetadata
Key Information
- Industry
- Data Management
- Founded
- 2019
Frequently Asked Questions
What is Amundsen?
Amundsen is an open-source metadata discovery platform created by Lyft for finding and cataloging data assets. It indexes tables, dashboards, and resources, then ranks results by usage patterns using a PageRank-style algorithm. It's hosted by the LF AI & Data Foundation.
How does Amundsen's search algorithm work?
Amundsen uses a PageRank-style search algorithm that ranks data results by usage patterns. Frequently queried tables surface first, improving discovery for organizations with large data warehouses. This approach helps analysts and engineers find relevant data faster than traditional keyword-only searches.
Which data warehouses does Amundsen integrate with?
Amundsen integrates with major data warehouses including Snowflake, BigQuery, Redshift, Hive, PostgreSQL, and MySQL. These integrations enable organizations to index and discover data assets across multiple warehouse platforms within a single centralized interface, simplifying metadata access for analysts and data engineers.
Is Amundsen still actively maintained?
Amundsen's development slowed significantly after 2023, when Stemma—a commercial version created by co-founder Mark Grover—was acquired by Teradata. The most recent software release was August 2024. The project remains incubation-stage at LF AI & Data with no published roadmap.
What companies use Amundsen?
Major organizations using Amundsen include Lyft, Square, Instacart, Snap, ING, Workday, Asana, and Brex. The platform serves over 30 organizations in total, primarily data teams at mid-to-large technology companies with mature data warehouses and complex metadata discovery and governance needs.
How does Amundsen compare to DataHub or OpenMetadata?
Amundsen's architecture is simpler and lighter than DataHub, enabling faster self-hosted deployment with lower operational overhead. However, DataHub and OpenMetadata are actively developed and offer advanced features like column-level lineage and sophisticated governance. For new deployments, community consensus favors more actively maintained alternatives.
How Amundsen compares
Direct head-to-head against 3 competitors. Picked by 7wData.
Amundsen
- Positioning
- Open-source metadata search and discovery platform for finding and cataloging data assets across organizations.
- Customer segments
- Data teams at mid-to-large tech companies running mature data warehouses
- Strengths
- Simple, lightweight architecture with fewer dependencies than DataHub, enabling faster self-hosted deployment
- Watch for
- No active development since August 2024; project maintainers moved to Teradata after Stemma acquisition in 2023
DataHub
- Positioning
- Open-source metadata and context management platform backed by Acryl Data; adopted by Apple, Netflix, Visa, and CVS Health.
- Customer segments
- Large enterprise data engineering teams needing actively maintained, AI-ready metadata discovery and governance at scale.
- Strengths
- Largest open-source metadata community (13,000 members); native lineage, observability, and governance in a single platform.
- Watch for
- Self-hosted deployment requires Kubernetes expertise; advanced AI governance features gated behind paid DataHub Cloud tier.
- Recent moves
- $35M Series B led by Bessemer Venture Partners (May 2025), targeting AI governance and context management expansion.
OpenMetadata
- Positioning
- Actively maintained open-source metadata platform; commercial backer Collate ships weekly releases with 120-plus connector support.
- Customer segments
- Data teams migrating from dormant catalogs; cloud-native organizations wanting self-hosted lineage and governance without vendor lock-in.
- Strengths
- Native column-level lineage out of the box; weekly release cadence; 120-plus connectors vs. Amundsen's narrower integration set.
- Watch for
- Commercial backer Collate raised only $10M Series A (July 2025); enterprise support capacity limited vs. better-funded competitors.
- Recent moves
- Collate raised $10M Series A (July 2025) to fund AI agent integration and enterprise go-to-market scaling.
Atlan
- Positioning
- Commercial managed data catalog and governance platform; Gartner 2025 Magic Quadrant Visionary for data and analytics governance.
- Customer segments
- Enterprise analytics and governance teams at Snowflake, Databricks, and dbt-heavy organizations needing managed catalog with compliance controls.
- Strengths
- Gartner-ranked #1 in lineage and impact analysis (2025); AI Governance Studio; Metadata Lakehouse for cross-tool context sharing.
- Watch for
- $34.9M ARR against $170M-plus raised as of March 2025 prompts enterprise procurement teams to flag long-term viability.
- Recent moves
- Launched App Framework with 21 technology partner integrations (August 2025) to extend governance across AI-native enterprise stacks.
Sources
- www.amundsen.io — Customer list (ING, Square, Workday, Snap, Cameo, Cimpress, Edmunds, EverFi, iRobot, LMC, Merlin, Remitly, Workday), core feature descriptions, and productivity claims
- github.com — Project architecture, components, and GitHub activity; confirmed 4.8k stars, 965 forks, 2,711 commits; support for Snowflake, BigQuery, Redshift, PostgreSQL, MySQL
- lfaidata.foundation — LF AI & Data Foundation incubation-stage status; contributed by Lyft in July 2020
- atlan.com — Project dormancy, latest release August 2024, no active maintainer since Stemma acquisition by Teradata in 2023, failed to graduate from incubation, competitive positioning against DataHub
- developer.squareup.com — Square as a verified customer implementing Amundsen for column-level access control
- www.castordoc.com — Architectural overview, Neo4j/Elasticsearch stack, feature descriptions