Introducing Commons

Today, we’re introducing Commons, data coordination infrastructure for organizations that work with each other’s data.

Commons connects the work of collecting and sharing data with how people find and use it. Publishers share data from their existing systems. Other organizations subscribe to local copies they control and receive published updates with a history they can inspect. Capabilities are reusable extensions that define how a dataset accepts contributions and how people or applications work with its contents.

We’re building Commons so useful data can reach more people and the effort spent preparing it can give the next team a better starting point. Our ambition is to make Commons the infrastructure people turn to for data coordination across ecosystems.

Why we’re building Commons

Sharing data creates work that continues long after the first exchange. For a publisher, another program or partner can mean another delivery process to maintain. For the team using the data, it can mean another import to repair or another job that checks for updates. When a file is replaced or a link disappears, finding the source behind an earlier report can become a project of its own.

That work takes care and judgment. Someone has to check the records and explain what a field means. A team may spend considerable effort preparing a dataset, only for another team to repeat the work because it has no easy way to use what the first team prepared.

Public data makes this need especially clear. Government bodies, research collaborations, and industry networks depend on information contributed, maintained, and used by different organizations. We’re building Commons to give those organizations a shared way to work with data as it changes, while they keep their own systems and responsibilities.

Collection starts with common rules

Coordination begins before a dataset is ready to publish. A program collecting contributions needs agreement on what a record should contain and which checks it must pass. Its participants may use different applications and supply information through files, APIs, or streams.

With Commons, an ecosystem can put that collection model into a shared capability. An ingestion capability reads a supported source and maps its records into the agreed structure. Validation checks each record as it is written. Participants can contribute through suitable integrations while continuing to use their existing tools.

Program teams can apply the same requirements across contributions, and contributors can prepare data against a clear set of rules. The people who understand the domain still decide what the information means. Commons helps them apply those decisions across participating datasets.

One publication can serve many organizations

Once data is ready to share, a publisher chooses what to publish and who can access it. Authorized organizations subscribe to the same maintained publication, even when they belong to different programs or networks. Published changes reach their local copies through Commons.

For a publisher, this reduces the need to manage a separate delivery process for every relationship. Subscribers query and process the copies they hold, so they don’t need to send every request back to the publisher. Teams can use the same local data in reporting, operational applications, and AI, or feed a data lake through an API integration.

A local copy lets a team continue working with the data it has received during a connection interruption. The application decides whether that version is recent enough for the task. Retained history also helps teams revisit the inputs behind an earlier report, provided the application records which versions it used. Signatures let subscribers check who signed a published change and whether its content has been altered.

A subscriber can become a publisher too. If your team has cleaned, combined, or enriched data, you can publish the prepared result for others to use. You decide when to refresh it and publish a new version. The next team gets a maintained source it can build on, and the preparation you did can benefit more than one project.

Help people find what they need and use it

Receiving a dataset is only part of making it useful. People need to find the relevant records and work with them in a form suited to the task. A researcher may want a table to explore; an application may need a query that returns a particular set of records.

Commons capabilities provide table views, full-text search, and queries over the local data. They can also give an AI assistant tools to list, read, and search a dataset through the Model Context Protocol, or MCP. The same shared information can support different ways of working.

Wider network discovery is on our roadmap. We want people to find a shared source, understand what it contains and how to access it, then subscribe to its published updates.

Capabilities let each ecosystem shape Commons

Different ecosystems need different rules and ways to work with data. An accreditation network may need shared rules for the certificate information its members publish. A research collaboration may need a different collection model. We designed capabilities so those needs can become reusable parts of Commons.

A capability is a small, signed package of rules and tools for a dataset. Developers can define how data comes in, which records are accepted, and how people and applications read it. A package can combine an importer with validation rules and a searchable view, then be published as a versioned release for other datasets to use.

This lets a team add support for a source format or a new way to use data without rebuilding the underlying storage and synchronization. An ecosystem can share the rules it needs while participants choose their own views and applications. Data preparation can be reused through a publication; the behavior that makes it useful can be reused through a capability.

Catalogs help people locate sources, APIs provide access, and data platforms support preparation and analysis. Commons focuses on coordinating data across these independently operated systems. Shared collection rules and maintained subscriptions give participants a common basis to work together, while capabilities let them adapt it to their own needs.

Explore where Commons could fit

We’re excited to see what people build when they can spend less effort coordinating the same data and more time putting it to use.

Think about one dataset your team maintains, collects, or depends on. Where does the work start to repeat? That is a useful place to begin exploring Commons.

We look more closely at what happens after a dataset is shared in When public data changes.

Explore how Commons works and see what capabilities make possible. If you can see a place for it in your work, join the waitlist to try Commons. We’ll share product updates and let you know when access becomes available.