OpenAI recently made waves when it announced that it would start using the Model Context Protocol (MCP). This open-source standard is changing the game for developers integrating AI models with data. MCP allows developers to create seamless, two-way experiences connecting any data source to AI-powered business tools like chatbots and virtual agents. This added functionality improves the accuracy of generated AI responses by giving access to wider, more pertinent data. This new development comes as a huge victory for OpenAI. It recently built MCP support into every one of its products, including the new ChatGPT desktop app.
What makes MCP truly unique is that it offers a framework for these AI assistants to interface with systems where data actually resides. MCP gives AI models the ability to access various data sources. These range from business applications to content storage infrastructures to application development ecosystems. This unique capability helps AI models conduct tasks with much higher efficiency and accuracy. Now, developers can define “MCP servers” to expose structured data in a machine consumable way. They can produce “MCP clients”— apps and workflows that tap into these servers whenever required.
In just a few years, the protocol has turned into a rich open standard with increasing backing from cities, and thousands of active integrations. Supporting MCP Companies such as Block, Apollo, Replit, Codeium and Sourcegraph have already added MCP support to their platforms. OpenAI’s adoption is just the latest development that cements MCP’s role at the heart of the industry.
"People love MCP and we are excited to add support across our products."
From: OpenAI CEO Sam Altman
OpenAI just announced they are adding the integration of MCP to their Agents SDK. In just a little bit, you’ll start to see support for the ChatGPT desktop app as well as the Responses API! This decision is part of a larger trend towards implementing large language models (LLMs). It works easily with the data and software infrastructures that already exist.
"Excited to see the MCP love spread to OpenAI – welcome! 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."
From: Anthropic chief product officer Mike Krieger