A Better Way to Give Credit: Introducing the Wikimedia Attribution Framework and API

For 25 years, Wikimedia volunteers have built the largest collection of open knowledge in the world. Millions of people access that content every day, starting from readers worldwide engaging with knowledge directly on Wikimedia projects, to smart speakers, AI chats, and other new formats summarizing or repackaging our content. New channels expand access to knowledge, but where that content came from is often hidden or obscured. For people encountering information from a Wikipedia article in an LLM answer or a Commons image used in an indie game, it may be unclear how it was processed, or how to find more context about it.

While new ways for people to gain knowledge are welcome, this shift in visibility also poses a challenge to the sustainability of Wikimedia projects. Without visibility, the pathway for new readers to become the next generation of contributors and donors is lost, ultimately affecting the mechanisms that ensure the continued vibrancy of the community and quality of the content.

To address these challenges, the Wikimedia Foundation is releasing two tools: the Wikimedia Attribution Framework V1 and the Attribution API [beta]. These tools make it simple for developers to fairly credit volunteer contributions and ensure a healthy  free knowledge ecosystem. When anyone encounters Wikimedia content, we want them to know that it comes from our projects, and they are invited to participate.

Read on to learn more about what they are, why they matter, and how you can help shape their next steps.

Why Attribution Matters: A Symbiotic Ecosystem

Attribution isn’t just about credit; it’s about sustaining the free knowledge ecosystem. The changing trends around direct traffic to Wikimedia websites signals the shift to a new paradigm where access to Wikimedia content is more diverse and distributed. As more content is consumed in external environments beyond the wikis, we want to help readers recognize Wikimedia content anywhere and motivate them to participate in the free knowledge ecosystem.

We want good attribution practices to be as easy to apply as possible, to help reusers seamlessly invite their audiences in different contexts to engage with Wikimedia content (through deeper reading, donating, or editing) before they even visit our sites. With clear visibility and greater recognition of Wikipedia’s ubiquitous value, more people will continue to participate and keep the healthy, virtuous cycle of high quality, free knowledge for everyone.

Introducing the Attribution Framework V1

The Wikimedia Attribution Framework sets guidelines on how to provide sustainable attribution when reusing Wikimedia content. The framework is designed to be flexible. This makes it possible to apply it to different reuse scenarios (such as Search, AI Assistants, Social media and more) while providing a consistent experience overall.

All reuse scenarios are based on a catalog of standardized signals. These have been organized in different levels:

• Essential Attribution: The foundational elements required to meet the obligations set by the relevant open licenses, while keeping off-wiki readers aware of the original communities that created the content. These signals make it straightforward for developers to satisfy licensing requirements for different content types, including clearly indicating whether content has been summarised or otherwise changed by the reuser (instead of being cited verbatim). 
• Trust & Relevance. Signals that highlight the credibility and living nature of Wikimedia content, while piquing readers’ curiosity and interest. Contributor and reference counts reveal the depth of collaborative, human effort behind each article, communicating that many people worked together and have sources to back it up. Additional signals like pageviews, last-updated timestamps, and trending indicators further reinforce the relevance and vitality of the content.
• Ecosystem Growth. Active interventions inviting users to participate. Providing additional info and call to actions (CTAs) readers can take to be more involved in the open knowledge ecosystem. Inviting readers to “Create an account”, “Download the app”, or “Donate” can be relevant ways to participate depending on the context.

The Attribution Framework compiles best practices for the attribution of Wikimedia content, making it easy for external reusers to apply them to their particular case. Organizations and developers interested in applying the framework to a specific context can find the documentation and the visual representations that illustrate how attribution can be provided in specific scenarios within the Attribution Framework. Guidance is defined in a flexible way, with multiple options to adjust the scenarios to the particular context using configurable examples. 

The initial release of the framework represents an early beta version, which will evolve over time with more use cases, signals and CTAs. We want to learn from external reusers, developers, and interested members of the Wikimedia community, and improve the framework based on their feedback.

The Developer Solution: The Attribution API (Beta)

Attribution considerations are relevant to a broad audience of reusers of Wikimedia content. Those making games, offering search services, using project content for research, building alternative reading experiences, or contributing to anything else happening off-wiki, probably need to properly attribute Wikimedia content.

However, this task has not always been easy for developers. In the past, external developers building products based on Wikimedia content had to stitch together complex requests or parse raw HTML to meet license requirements and display rich attribution signals. These barriers often led to Wikipedia content being reused with sub-optimal attribution.

The Attribution API (currently available as a beta module) complements the guidelines provided in the Attribution Framework to make it easy for developers to obtain the data they need to properly attribute Wikipedia articles and media files from Commons. It abstracts the complexity behind a simple endpoint that returns exactly the information required by the framework’s signals on a per-page basis. Reusers may also flexibly filter and target the specific signals relevant to each scenario or product need. As we continue to invest in this space, we expect to offer additional endpoints for project-level attribution, easier ways to get attribution information for the images embedded in articles, and more. Upcoming and recent changes can be tracked on the Attribution API project page.

NOTE: Although the attribution framework is designed for everyone, this specific API is primarily intended for mission-supporting users and use cases. Wikimedia Enterprise will offer similar information in their structured responses for high volume commercial reusers, who are expected to follow the same Attribution framework guidelines. For more information about all of the options for retrieving attribution signal data, see the technical implementation section of the Attribution Framework.

Participate!

The Attribution Framework and API are at an early stage. They have been released to start learning from different reusers in different situations, and will evolve based on your feedback.

You can read the attribution framework guidelines to review the recommended practices for attribution. Learn more about the project on the project page, and reach out if you’re interested in being featured as an early adopter or to share any thoughts. 

You can also explore the Attribution API in the REST Sandbox, which is currently available on all Wikimedia Foundation hosted wiki projects, such as English Wikipedia or Meta-wiki, follow along on the project page, and give us feedback on your experience.