Thought Leadership
OpenAI is taking another step towards developer-friendliness. The company has published the so-called “Responses API”, which provides developers with the building blocks they need to develop their own AI agents.
These agents are able to search the internet, analyze files and perform tasks on a computer independently. “There are some agents that we will be able to build ourselves, like Deep Research and Operator,” explains Olivier Godement, Head of Product for the OpenAI platform, in an interview with The Verge. “But the world is so complex, there are so many industries and use cases… and so we’re super excited to provide those foundations, those building blocks for developers to build the best agents for their use case, their needs.”
Responses API: Versatile tools for developers
The Responses API is intended to provide several powerful functions. These include a web search tool based on the same model that ChatGPT uses for its search function. Developers can use it to retrieve real-time information and source information from the Internet – the GPT-4o and GPT-4o mini models are supported. Another feature is the computer use option, which is based on the OpenAI operator model and enables agents to perform tasks on behalf of a user. The API also provides a tool for searching through large volumes of documents. OpenAI sees potential applications here in customer support, for example, to efficiently search FAQs, or for legal assistants who need to research previous cases.
Agents SDK: Orchestration of complex tasks
In addition to the Responses API, OpenAI has also announced the Agents SDK. This software development kit is designed to help developers orchestrate the workflow of AI agents. “The Responses API is like this atomic unit of using models and tools to do a particular thing,” Nikunj Handa, Product Manager in OpenAI’s API team, explains to The Verge. “The Agents SDK is having multiple of those atomic units work together to solve even more complicated tasks.” This should make it easier for developers to manage multiple agents and ensure that they work towards a common goal.
The introduction of the Responses API and the Agents SDK builds on OpenAI’s existing tools, such as the Chat Completions API, which allows developers to create AI tools that respond to user requests. OpenAI also plans to replace the existing Assistants API with the Responses API in mid-2026. According to the company, “significant improvements” have been integrated into the new API based on feedback from developers on the Assistants API.
