
Language models use news articles and publishers as sources without permission and often return inaccurate responses, while the newsrooms that produced the information are left out of the equation.
Developed by Ambiental Media, Jor-MCP is an infrastructure created to confront this. The beta version code is already available on GitHub.
A server that allows newsrooms to make their content available to artificial intelligence systems, under their own rules.
Today, journalistic content is scraped without permission. With Jor-MCP, publishers will have the opportunity to offer structured and authorized access to AI applications and agents. The newsroom defines what goes into the server.

Jor-MCP is built on the Model Context Protocol (MCP), an open standard proposed by Anthropic and now adopted by companies, public institutions and research organizations around the world.
Jor-MCP is one of the first applications of the protocol aimed at journalism and the first in Brazil.
AUTHORIZATION
Each newsroom decides which content is available to AI systems and under what conditions. What today only happens as informal extraction also gains a consensual access path.
GOVERNANCE
AI systems that want to access journalistic content connect through the newsroom's own server. With this, the newsroom gains visibility into data such as access frequency and volume.
POTENTIAL COMPENSATION
With authorization and visibility, the technical conditions are created for compensation models to begin to be explored — something currently unfeasible when content circulates without clear identification.
Each publisher maintains its own Jor-MCP server, hosting a structured version of its content.
AI applications — assistants, research agents, fact-checking tools — connect to these servers to query information directly at the source, instead of relying solely on what the model absorbed during training.
The result: potentially more accurate responses for the user, control for those who produce and a path for journalism to participate in the AI infrastructure — even though, alone, Jor-MCP does not eliminate the asymmetries with large platforms.

Jor-MCP will be published as open source: code, documentation and learnings will be public, so that any outlet can implement it independently. To pilot and maintain the solution, your newsroom will need its own technical team.
To pilot and maintain the solution, your newsroom will need its own technical team.
If you lead a journalism organization or work with technology in the field, leave your contact to receive updates about the launch, documentation and testing opportunities.
List used only for project updates. No spam.
Ambiental is a Brazilian journalism organization that operates at the intersection of science, data and technology. Our reporting transforms complex information into accessible and engaging content, focused on socioenvironmental and climate issues, and has been recognized by awards such as Sigma Awards, WAN-IFRA and Cláudio Abramo.
We use artificial intelligence in stages of journalistic work and have developed initiatives such as Capí, an AI specialized in climate change.
Ambiental is the only Brazilian organization to join the international Dirty Data consortium, which investigates the impacts of data center expansion. Jor-MCP is born from this trajectory.
Produced by:
Supported by: