OpenAI is adding support for MCP, an open-source technology that uses large language models (LLMs) to perform tasks in external systems.
OpenAI CEO Sam Altman announced the move this week, SiliconANGLE reports. This development is special, partly because MCP was developed by Anthropic PBC, the ChatGPT developer’s best-funded startup rival.
At launch, support for MCP is available in OpenAI’s Agents SDK, an open-source toolkit for developers to build artificial intelligence agents. In the coming months, OpenAI will integrate MCP into ChatGPT’s desktop client and the Responses API, the application programming service that gives developers access to the LLMs.
Accelerating integrations
Companies can make their LLM’s more valuable by connecting them to external systems. For example, a retailer can give a language model access to a product listings database and use it to generate shopping advice. Building such integrations used to be a time-consuming job.
Anthropic’s MCP protocol facilitates this task. It offers software building blocks for developers to connect their LLMs to external systems quickly. According to the company, creating integrations with the protocol takes less than an hour in some cases.
MCP enables LLM not only to retrieve data from external systems but also to perform actions in those systems. An LLM optimized for coding tasks could use the protocol to execute a configuration script on a cloud instance. An AI-driven marketing tool could, in turn, enter advertising performance in an analytics application. OpenAI’s decision to add support for MCP makes such functions available to users of ChatGPT Desktop, Responses API and Agents SDK.
MCP uses JSON-RPC technology to move data between LLMs and the systems they connect to. According to Anthropic, the latest protocol version adds a feature known as JSON-RPC batching. This allows MCP to package multiple LLM data requests into one large request, increasing efficiency.
The new version also makes it easier for MCP-enabled systems to send notifications to the LLMs that have access to them. In addition, Anthropic has upgraded MCP’s authorization mechanism to OAuth 2.1, the latest version of OAuth, a technology that helps set up secure connections between applications.
Microsoft’s own integration
Against these updates, OpenAI investor Microsoft launched a new MCP integration of its own. The company released a tool called Playwright MCP, which combines the protocol developed by Anthropic with its own Playwright software.
Microsoft originally developed Playwright to facilitate website bug testing. The software, available under an open-source license, can automatically execute actions in a browser. Microsoft’s recently released Playwright MCP tool utilizes Playwright’s web browser functions to allow LLMs to interact with web pages.
Developers can use the tool to automate tasks such as filling out forms. According to Microsoft, Playwright MCP also enables LLMs optimized for coding to automate website testing tasks.