Open Agentic Web Challenge offers $13,000 for interoperable AI agents
A global group of technology, research, standards and civil society organizations has launched the Open Agentic Web Challenge, offering $13,000 for open and interoperable AI agent projects. The first competition is open on DevPost until January 11, 2027.
- The Open Agentic Web Challenge offers a $13,000 prize.
- Entries are open on DevPost until January 11, 2027.
- The challenge uses existing Web technologies including HTML and JavaScript.
- Backers include Cognizant, MIT Media Lab, Mozilla, AWS and ITU-T Study Group 17.
- Winners will be announced at the World Economic Forum meeting in Davos in January 2027.
A competition for an open agent network
Announced at Web Summit in Lisbon on November 13, 2026, the initiative asks developers, researchers, start-ups, companies and students to create publicly accessible, interconnected agentic websites. Participants can also develop technologies and standards that allow independently built agents to discover one another, establish trust and work together safely.
The prize will recognize the strongest contributions, with winners scheduled to be revealed at the World Economic Forum’s annual meeting in Davos in January 2027. The organizers are also inviting additional organizations to join as partners.
Why the organizers say open standards matter
AI agents are beginning to find online services, communicate, negotiate and conduct transactions for people and organizations, according to the announcement. The infrastructure for agent discovery, identity, trust and the exchange of value is still being developed.
The organizers warn that personal agents are currently largely restricted to the platforms where they were created. Without open standards, that separation could make it harder for individuals and businesses to access services. The challenge instead explores a model in which people can choose their agents, organizations can make their services available to agents, and agents from different developers can interact across platforms.
Rather than requiring an entirely new network, the competition will use the existing Web and common technologies such as HTML and JavaScript. Projects can test practical approaches to discovery, identity, multiagent orchestration, portable agent programs, transactions and governance.
Tools and organizations supporting the challenge
Participants can build their own components or use open source starting points and research from participating organizations. The MIT Media Lab’s NANDA Web Agentica and NandaIndex provide research prototypes for signed, portable agent packages that run in a user controlled environment, such as OpenClaw or a browser.
Cognizant’s Neuro San is an open source framework for building and orchestrating multiagent systems. Mozilla’s Otari is an open source large language model router designed to give agentic applications flexible access to different models. Using these projects is optional.
The initiative brings together Cognizant’s AI Lab, the MIT Media Lab’s NANDA program, Mozilla, AI Commons, the International Telecommunication Union’s Study Group 17, Amazon Web Services and the University of Texas at Austin’s Computer Science Department. Their stated focus areas include multiagent systems, open Web principles, agent identity and trust, knowledge traceability, security standards, cloud infrastructure and user control.
For users and organizations, the effort could shape how agents interact with online services if the prototypes and standards developed through the challenge prove practical. The announcement’s emphasis on open protocols, permissions, inspectable interactions and traceability points to a Web where people can select agents from different providers and organizations can participate without relying on a single closed platform.

