May 22nd.OpenAI The company published a blog post yesterday, May 21, announcing the expansion of the Responses API,Remote connection support MCP Server, Image Generation, Code Interpreter, and File Search Optimization to help developers build smarter smart body applications.
1AI querying the public data, the Responses API is OpenAI's stateful API that supports a variety of new tools including web search, file search, and computer usage, providing developers with a cleaner, more flexible way to interact with OpenAI models.
In this extension to OpenAI, the Responses API supports remote connections to MCP servers, image generation (gpt-image-1 models), Code Interpreter, and improved file searching, with full support for the GPT-4o series, the GPT-4.1 series, and the OpenAI o-series of inference models.
Among the most noteworthy improvements.It supports remote connection to MCP servers.. With just a few lines of code, developers can enable OpenAI models to connect to popular MCP server tools such as Cloudflare, HubSpot, Stripe, and more, easily accessing data sources that users rely on.
The new Image Generator tool supports real-time streaming previews and multiple rounds of editing, allowing developers to fine-tune images step-by-step; the Code Interpreter tool facilitates data analysis, complex math, and coding problem solving; and the File Search tool supports multi-vector storage searches and attribute filtering.
The Responses API also adds a background mode that supports asynchronous processing of long tasks to avoid timeout issues; an inference summary feature that summarizes the model inference process in natural language for debugging and optimizing the user experience; and encrypted inference items that ensure that Zero Data Retention (ZDR) users can reuse the inference items without storing the data, reducing cost and latency.