Jor-MCP
Jor-MCP is an open infrastructure developed by Ambiental to enable news organizations to make their content available to artificial intelligence systems in a structured and transparent way, under their own rules. Based on the open Model Context Protocol (MCP) standard, the project has the potential to offer an alternative to data scraping by AI models, creating new paths for governance, attribution, access control, and the economic sustainability of journalism.
How It Works
The growing adoption of artificial intelligence models has transformed the way news is consumed and reused. Today, AI assistants consult news reports to answer questions, generate summaries, and power autonomous agents. In many cases, however, this process takes place without clear mechanisms for attribution, licensing, or compensation for the news organizations that produced the original content.
Jor-MCP was created to address this challenge. Instead of allowing AI systems to access web pages filled with menus, advertisements, scripts, and other unnecessary elements, the server delivers only the requested content in a structured format designed for machine consumption. This reduces processing and token usage, improves the contextual accuracy of responses, and makes it possible to include metadata such as authorship, source, licensing, and other information defined by the news organization itself.
The project was created within Ambiental, an independent organization specializing in investigative journalism on climate and the environment, through collaboration between journalists and its technology team. Jor-MCP receives financial support from Codesinfo – Innovation Fund for Journalism, a program of Projor – Institute for the Development of Journalism, sponsored by the Google News Initiative, as well as from the JournalismAI Innovation Challenge, a project of Polis LSE, the journalism think tank at the London School of Economics and Political Science.
Documentation
Jor-MCP is implemented as a server compatible with the open Model Context Protocol (MCP) standard, allowing artificial intelligence agents to query journalistic databases directly without having to interpret complete HTML pages. Its architecture was designed to be lightweight, stateless, and composed of a limited set of tools, reducing infrastructure costs and making it easier for small and medium-sized journalism organizations to adopt.
In its current version, the server offers three main operations: searching content published by the newsroom on its website and data repositories; retrieving the clean and structured text of a specific article; and listing the most recent publications, allowing AI agents to stay up to date with newly published content.
The project follows an open and replicable architecture. All code is available as open source under the Apache 2.0 license, together with technical documentation so that other organizations can deploy their own servers and independently define their operating rules. The strategy is designed so that each news organization maintains its own infrastructure, access controls, and usage policies, avoiding dependence on a centralized platform.
More than a finished product, Jor-MCP represents an open infrastructure under development. Its goal is to help ensure that journalism takes part in defining the technologies that will mediate the flow of information between newsrooms and artificial intelligence systems, strengthening principles of transparency, governance, public interest, and sustainability across the journalism ecosystem.
