OpenAI is moving to support Anthropic's Model Context Protocol, a standard designed to help AI assistants connect with the systems where useful data already lives. The decision matters because MCP is meant to make AI responses more relevant by giving models a structured way to work with business tools, software, content repositories, and app development environments.
OpenAI is adopting a rival-backed standard
OpenAI CEO Sam Altman said in a post on X on Wednesday that OpenAI will add MCP support across its products. The rollout includes the desktop app for ChatGPT, with MCP already available in the Agents SDK and support for the ChatGPT desktop app and Responses API coming soon.
Altman framed the move as a response to developer interest. “People love MCP and we are excited to add support across our products,” Altman said. “[It’s] available today in the Agents SDK and support for [the] ChatGPT desktop app [and] Responses API [is] coming soon!”
The important point is not just that OpenAI is adding another integration feature. It is that OpenAI is embracing a standard created by Anthropic, a rival in the AI assistant market. That signals a practical recognition that model usefulness increasingly depends on access to the right context, not only on the model itself.
What MCP is designed to do
MCP is an open source standard for connecting AI models to data sources. Its purpose is to help AI systems produce better and more relevant answers when a request depends on information stored outside the model.
The source article describes MCP as a way for models to draw from places such as business tools and software, content repositories, and app development environments. In practical terms, that means an AI-powered application can be connected to the systems a user or company already relies on, rather than being limited to a standalone chat window.
MCP also supports two-way connections between data sources and AI applications. Developers can expose data through “MCP servers” and then build “MCP clients” such as apps and workflows that connect to those servers on command.
That client-and-server model is central to why MCP has attracted attention. It gives developers a common structure for linking AI-powered applications, including chatbots, to outside systems. The value is clearest when an AI assistant needs current or specific information from a tool, repository, workflow, or development environment.
Why developers and platforms are paying attention
Since Anthropic open sourced MCP, several companies have added support for the standard. The list in the source includes Block, Apollo, Replit, Codeium, and Sourcegraph.
That growing platform support gives OpenAI's decision more weight. A standard becomes more useful when multiple developers and companies build around it. If more tools expose data through MCP servers and more applications act as MCP clients, the protocol can become a shared layer between AI assistants and external systems.
Anthropic chief product officer Mike Krieger welcomed OpenAI's move in an X post. “Excited to see the MCP love spread to OpenAI – welcome!” he said. “MCP has [become a] thriving open standard with thousands of integrations and growing. LLMs are most useful when connecting to the data you already have and software you already use.”
That statement captures the broader direction of the AI application market. The next step for assistants is not only answering general questions. It is helping users complete tasks in the tools, repositories, and software environments they already use.
What changes for OpenAI products
OpenAI has said MCP support is available today in the Agents SDK. The ChatGPT desktop app and Responses API are next on the stated path, with support coming soon.
For developers, the Agents SDK is the first concrete place to look. For users of ChatGPT's desktop app, the change is still ahead. For builders working with the Responses API, OpenAI has also indicated support is on the way.
The source article does not provide a full rollout schedule or detailed product behavior. It says OpenAI intends to share more about its MCP plans in the coming months. Until then, the confirmed facts are limited to the current Agents SDK availability and the coming support for the ChatGPT desktop app and Responses API.
Still, the direction is clear: OpenAI wants its products to work with a protocol that is already gaining support outside its own ecosystem. That could make it easier for developers to build AI workflows that are less isolated and more connected to the systems where work already happens.
The larger signal for AI assistants
The adoption of MCP points to a simple but important shift. AI assistants become more useful when they can reach relevant information and software, especially when a user's request depends on context that is not inside the model.
That is why this move is notable even without a long list of new product details. MCP is about connecting models to data sources and applications. OpenAI's support brings that idea into its own product ecosystem while acknowledging a standard that began with Anthropic.
The result is a more interoperable path for AI-powered apps, workflows, and chatbots. Developers still need to build the MCP servers and MCP clients that make those connections work, but OpenAI's support gives the standard a larger role in the AI development landscape.