OpenAI Launches Operator: New API for Custom AI Agents

OpenAI has unveiled a new API designed for developers to create and integrate their own AI agents, similar to the company's existing offerings such as the general-purpose Operator and the research-focused Deep Research agent. This new toolset includes the Responses API, integrated tools, and a Software Development Kit (SDK) aimed at enhancing user experience in AI development.

The Responses API serves as a combination of the previously available Chat Completions API and the Assistants API. The Chat Completions API is responsible for facilitating communication, while the Assistants API allows the agent to execute tasks beyond mere conversation. This comprehensive API features integrated tools such as Web Search, enabling agents to access the internet, File Search for handling large documents, and Computer Use, which permits operational capabilities on local machines. However, for the Web Search functionality to be effective, developers must utilize either gpt-4o or gpt-4o mini.

As part of its commitment to continuous innovation, OpenAI emphasizes that this is just the initial set of resources provided to developers. The new SDK is designed to orchestrate both single-agent and multi-agent workflows, incorporating integrated observability tools. These tools are intended to enhance the manageability and traceability of agent workflows, allowing developers to monitor and analyze their operations more effectively.

The Responses API is now available at no additional cost for users with a developer account. The standalone Chat Completion API will remain accessible; however, OpenAI recommends that new integrations utilize the Responses API, even if the full capabilities of agent functions are not currently required. The Assistants API is set to be phased out, with OpenAI anticipating its complete discontinuation by mid-2026, at which point the Responses API will fully take its place.