Collibra

Overview

Collibra is an independent governance platform, source-agnostic by design: it holds metadata about data it does not store, spanning cloud platforms, databases, enterprise applications, BI tools and legacy systems.

The capability that sets it apart from the other governance platforms is that it actually enforces policy rather than only documenting it. Protect pushes grants and column masking down into AWS Lake Formation, BigQuery, Databricks and Snowflake.

Importantly it enforces natively in those platforms rather than sitting in the connection path as a proxy, so there is no new hop for queries to travel through.

Connectivity runs two ways. Edge is a run-time you host on your own Kubernetes close to the data; Collibra Cloud sites are the hosted equivalent if you would rather run nothing.

It is modular to a fault, with 26 named products from business glossary and catalogue through policy manager, reference data, protect and workflow designer, which means scoping a purchase takes real work.

The marketplace is its consumer-facing face: a curated storefront of datasets, tables, feature sets, reports and APIs, each carrying descriptions, quality scores, certifications and lineage.

Notably it publishes that marketplace as serving AI agents as well as people, which is where the product is heading.

Policy expression is attribute-based, combining data protection standards that mask columns by classification with access rules that grant by role, and you can register your own masking function in Databricks or Snowflake.

Check enforcement details per platform, because they differ: in Lake Formation and BigQuery protection and access control are integrated, while in Databricks and Snowflake they are separate concerns.

Its 2026 direction is agentic AI governance, with an AI command centre launched in May 2026, an AI governance product and an agent platform in preview, backed by two acquisitions in 2025.

Key features and capabilities

The same headings are used for every governance & stewardship entry, so two tools can be read side by side.

What it catalogues

• Catalogues datasets across cloud platforms, databases, enterprise applications, BI tools and legacy systems
• More than a hundred native integrations published for source connectivity
• Metadata is harvested through Edge, which you host, or Collibra Cloud sites, which they host
• Data profiling and data samples are separate named features alongside the catalogue
• A browser extension surfaces catalogue context inside Power BI and Tableau

Ownership and stewardship

• Communities and domains are the ownership containers, established through a stewardship product
• Data products and data contracts are a published platform feature, with an agent for data products
• A marketplace presents curated datasets, tables, feature sets, reports and APIs as a storefront
• Each listing carries business descriptions, quality scores, certifications and lineage
• The marketplace is published as serving AI agents as consumers, not only people

Glossary, policy and classification

• A business glossary with configurable asset types, and a separate policy manager for digital policies
• Classification as a platform feature, plus assessments that score data risk from templates
• Policies are enforced, not merely documented, pushing grants and masking into the source platform
• Enforcement is native in Snowflake, Databricks, BigQuery and Lake Formation rather than through a proxy
• Custom masking functions can be written in Databricks or Snowflake and registered for use in a standard

Lineage and impact

• A dedicated lineage product visualising data flow end to end across systems
• Lineage is produced by dedicated harvesting service instances rather than as a side effect of cataloguing
• Column-level guarantees per source are not published
• Impact analysis comes through relationships held in the metadata graph

Workflow and collaboration

• A full workflow designer for modelling processes and forms
• A data helpdesk manages data issues and tracks progress
• Access requests route to the right stakeholders automatically, integrating with Jira and ServiceNow
• Collaboration through collections, ratings, reviews and certifications
• Access tracking produces an auditable trail of who accessed what and when

Integrations

• More than a hundred native integrations, plus outbound metadata push to Unity Catalog and Snowflake
• Policy enforcement targets Lake Formation, BigQuery, Databricks and Snowflake
• A documented developer portal with REST APIs and GraphQL, generated from shipped OpenAPI specifications
• A marketplace distributing listings and packaged solutions
• Named partner of the year awards from Databricks and Google Cloud

Where it runs and what it costs

• Software as a service managed by Collibra on AWS and Google Cloud, across all major regions of both
• Edge sites are customer-run, one per network, security domain, region or private cloud
• Self-hosted and on-premises deployment is not published
• Sovereign or government cloud is not published
• No unit of pricing is published

Pricing

Price on requestQuote only; nothing published

Entirely quote-only. Collibra publishes no prices at all: there is no pricing page, no editions, no per-user or per-asset list price, and no currency or unit of pricing anywhere. The documentation says only that your contract and assigned roles determine which capabilities you can access, which is a fair warning that scoping matters here given 26 separate products. The route to a number is a demo request or a call with a regional office, so budget time for a commercial conversation before you can compare it with anything.

Vendor pricing page →

Demos and videos

• What's new in the Collibra platform Webinar or talk · YouTube
• Seven steps to get started with Collibra Tutorial · YouTube

About Collibra

Collibra was founded in 2008 by Felix Van de Maele, its chief executive, and Stijn Christiaens, and runs dual headquarters in New York and Brussels with around a thousand staff. It is private, last disclosing a $250m Series G in November 2021 at a $5.25bn valuation led by Sequoia Global Equities and Sofina, a figure now five years old and worth treating as historical. It bought Raito and Deasy Labs in 2025, the latter for unstructured data. It reports more than 700 customers including over a hundred of the Fortune 500, naming TELUS and Red Sea Global.

Founded 2008 · New York and Brussels · collibra.com