OpenAI recently released a new suite of developer tools designed to make it easier to create AI agents that can perform complex tasks autonomously. The update, announced last week, introduces a response API, an open source proxy SDK, and built-in tools for web search, file search and computer control – all designed to simplify how AI Systems interact with real-world information and applications.
Openai describes these agents as “systems that perform tasks independently on behalf of users,” meaning they can perform multi-step processes (such as researching topics or updating databases) under human guidance. The company’s goal is to reduce barriers to developers and businesses deploying powerful AI-driven assistants, thereby increasing accessibility to advanced AI capabilities.
Response API: Simplify proxy interaction
At the heart of Openai’s announcement is the new reply API, which is a unified interface for building AI agents. This API combines the conversational capabilities of the OpenAI chat completion API with the tooling capabilities of the previous Assistant API. In fact, this means that a single API call can now handle complex multi-step tasks, which may involve calling various tools or knowledge sources.
Openai said that the development of agents is simplified by reducing the need for custom code and timely patching to simplify the API’s response API. “The response API is designed for developers who want to easily combine OpenAI models and built-in tools into their applications without the complexity of integrating multiple APIs or external vendors” The company explained in its announcement blog post. Previously, developers often had to orchestrate multiple API calls and carefully crafted tips to make AI agents do something useful, which was challenging and time-consuming. Using the new API, for example, a proxy can talk to the user, search for information through a web search, and then write a summary – all in one workflow.
It is worth noting that all developers can use the Response API for no additional charges beyond the standard usage costs. It’s also backward compatible: OpenAI confirms that it will continue to support its popular chat completion API for simple use cases, while the Senior Assistant API will be phased out in mid-2026 as its capabilities are collapsed into a responsive API.
Open Source Agent SDK Streamline Workflow Orchestration
The release also includes a proxy SDK, a toolkit that manages the workflow of one or even multiple interactive AI agents. With a notable move, OpenAI made this SDK open source, allowing developers and businesses to check code and even integrate non-Openai models into their proxy systems. This flexibility means companies can coordinate GPT-4 using OpenAI with another proxy powered by other AI models, all within the same framework.
The Adents SDK focuses on workflow orchestration – essentially tracking what an agent is doing and how to accomplish tasks. It is for example:
- Configurable proxy: Set up an AI proxy with predefined roles or specific tasks.
- Intelligent handover: A task based on multiple agents or between processes between contexts (e.g., another agent that collects data and then analyzes it).
- Security and security: Enable the agent to remain within certain ranges with input validation and content adjustment tools to prevent unnecessary output.
- Tracking and observability: Tools that gradually monitor and debug agent actions can help developers solve strategies and improve performance.
According to OpenAI, the toolkit can simplify complex use cases such as customer support bots, multi-step research assistants, content generation workflows, code review agents, or sales exploration automation. Through the open SDK, OpenAI also encourages community contribution and adoption in an enterprise environment where transparency and the ability of self-service host components are often important. Early adopters, including companies such as Coinbase and Box, have tried the Adents SDK to build AI-driven research and data extraction tools.
Built-in tools enhance AI capabilities
To make AI agents more functional in out-of-the-box functionality, OpenAI’s responsive API comes with three built-in tools that connect AI to external data and operations. These tools greatly extend what a proxy can do, not just generate text.
The built-in tools available at startup are:
- Web search: Allows AI agents to perform real-time web searches and retrieve the latest information, accompanied by a referenced source. This means that agents can answer questions using the latest news or facts on the internet and provide a reference for transparency. This tool is useful for construction agents like research assistants, shopping guides, or travel planners who require real-time information.
- File Search: Let agents quickly filter large amounts of documents or data provided by developers to find relevant information. This is actually a private knowledge base query tool – an agent can answer customer support questions by finding policy documents, or help with legal research by retrieving paragraphs from a library. The tool can be deployed in solutions such as a customer service robot or an in-house company assistant that requires reference to proprietary information.
- Computer use: A new feature (currently under research preview) that allows the AI agent to perform operations on the computer as if it were the human user operating the machine. Powered by OpenAI’s Computer Usage Agent (CUA) model, the tool translates AI’s intent into keyboard and mouse operations to navigate software, websites, or other digital interfaces. Essentially, it can automate tasks that are not easy for APIs – for example, entering data into traditional systems, clicking on a web application to test or checking information on a graphical interface.
By integrating these tools, AI agents can not only think through questions, but also take action – whether it means searching for information, retrieving specific data, or manipulating digital environments. This greatly expands the capabilities of the proxy and makes it more useful for real-world applications.
Openai envisions developers combining these tools as needed; for example, agents might use web searches to collect public information and file searches to extract internal data, and then use that combination of knowledge to draft reports or perform tasks. All of this can be curated through a unified response API instead of requiring separate services or manual integration.
Wide impact on AI adoption and accessibility
Analysts say the launch could accelerate the adoption of AI agents in the industry by reducing technical barriers. For enterprises, the appeal of these new tools is the ability to automate and scale processes without extensive custom development.
Today, using OpenAI’s building blocks, AI agents may handle routine tasks such as information retrieval, form processing, or cross-application data input (which may require a lot of encoding or multiple software systems). For example, the built-in search tool plugs the AI into its knowledge database or the web almost immediately, and the computer usage tool provides a way to interact with old applications without an API. At the same time, the open source nature of the agent SDK gives enterprises more control, allowing them to integrate these AI agents into existing infrastructure and even use different AI models as needed.
Openai’s move is part of a wider competition to enable developers to have agency building capabilities. Competitive tech companies and startups have been launching their own AI proxy platforms, and OpenAI’s comprehensive toolkit may help it stand out. In fact, while interest from global autonomous AI agents surges, time has come – for example, Chinese startup Monica has recently attracted attention for its agent’s feel, claiming it can outperform Openai’s own prototype agent in certain tasks. Through the open key parts of its platform and the provision of built-in tools, OpenAI appears to be responding to competitive pressures while also facilitating wider AI adoption.
From an accessibility perspective, these tools can democratize who can build advanced AI systems. Smaller companies and even individual developers may now find it feasible to create AI-driven assistants or workflows without the need for large research teams. Integration methods (one API call can handle multiple steps), and example availability in OpenAI documentation reduces the entry barrier for new immigrants. OpenAI also provides developers with an observability interface to track and check what agents are doing, which is essential for debugging and building trust in AI output. This focus on availability and security (using guardrails and surveillance) is expected to encourage more businesses to experiment with AI agents because they know they have supervision and control.
Artificial intelligence agents may become as common and essential as having the presence of the Internet. OpenAI’s latest tools can help bring that vision to life by enabling developers and organizational communities to build their own agents by making agent development more approachable.
