The first public edition of Dataville gives AI application developers access to fresh, attributed data from public sources and a selective set of licensed datasets through API or MCP—without relying on fragile scraping pipelines.
AI has made building software dramatically easier. Developers can now turn ideas into working applications faster than ever.
But building applications with trustworthy data remains surprisingly difficult.
Developers need to know where data comes from, how current it is, whether it is credible, and whether they have the right to use it.
For many developers, the workaround is scraping websites. That might be enough to get a prototype working, but it's a poor foundation for a production application. Websites change, scrapers break, information becomes stale, and attribution and usage rights can be unclear. What starts as a quick solution becomes an ongoing infrastructure and maintenance burden.
Today, we're launching the first public edition of Dataville to offer a better way.
Dataville connects knowledge creators with application builders and makes trustworthy data insanely easy to build with.
Building AI applications needs more than intelligent models
For applications involving financial analysis, research, compliance, business intelligence, or risk management, the quality of the underlying information matters as much as the intelligence of the model using it.
Dataville provides consistent access to fresh, attributed data from known sources, with provenance and applicable usage information preserved. This gives developers the information they need to evaluate the reliability, freshness, and permitted uses of the data in their applications.
The first public edition of Dataville brings together a selective set of licensed datasets and public sources, including Wikipedia, SEC EDGAR, the U.S. Census Bureau, arXiv, and other knowledge and research resources.
Wikipedia is an important piece of this first edition. Through Wikimedia Enterprise, Dataville provides developers access to structured, machine-readable Wikipedia content while preserving attribution and source context. This makes knowledge created by Wikimedia's global volunteer communities easier to incorporate into AI applications.
Through a single interface, developers can discover, query, and combine data from multiple sources including their own data. Dataville provides consistent access, through its API or its MCP server, preserving source and usage information across datasets while reducing the need to build and maintain separate integrations, scraping, ingestion, and normalization systems. For example, developers can build applications that use company filings and other public information to support financial research or monitor business risks.
Dataville's MCP server is now listed in the official MCP Registry under com.dataville/dataville-mcp, making it discoverable to developers building AI agents and applications using MCP.
Fred Gallagher, CEO, Dataville said, “Almost anyone with a good idea can now build an AI application. But building a dependable application is another matter. Developers should be focused on creating useful products, not worrying about whether the data behind them is accurate, current, or trustworthy. That’s the problem Dataville is solving.”
A healthier information ecosystem for AI
The information powering AI applications doesn't appear out of nowhere. It's created and maintained by researchers, journalists, publishers, analysts, and communities that contribute valuable knowledge.
Dataville is designed to connect these communities. Application builders get simpler access to useful data, while participating knowledge creators gain a new distribution channel without having to build and operate their own developer infrastructure. Participating creators control which data is available, how it can be used, and how it is priced. Dataville handles the infrastructure needed to put those decisions into practice.
Anil Dash, a technology entrepreneur, writer, and longtime advocate for a more open and equitable internet, sees the opportunity to create a more balanced relationship between knowledge creators and application builders.“Everybody knows that the arrival of the big AI platforms has brought a lot of frustrations: content creators feeling like they’ve gotten a bad deal, the people who create knowledge being treated as an afterthought, and developers constantly being sold a bunch of hype. Dataville makes me excited because it’s a real leap forward for the ecosystem, giving the next generation of applications a real shot at supporting everyone whose work makes AI innovation possible, while also fitting into the way that coders today actually work,” says Anil.
Start building with Dataville
Our ambition is to help more people build useful AI applications with data they can understand, evaluate, and trust.
The first public edition of Dataville is live, giving developers a simpler way to access and use trustworthy data in their AI applications.
Explore the available datasets, run your first query, and connect data to your applications through our API or MCP server.
We will continue expanding our data sources and capabilities to make trustworthy data even easier to build with.
Start building with Dataville for free — no credit card required.
