OpenAI is updating its Agents SDK to help enterprises build safer, more capable agents

Agentic AI is the tech industry’s newest success story, and companies like OpenAI and Anthropic are racing to give businesses the tools they need to create these automated little assistants. For this purpose, OpenAI has now updated the agents software development toolkit (SDK).introduces a number of new features designed to help businesses build their own agents that run behind OpenAI models.

New SDK capabilities include sandboxing capabilities that allow agents to run in managed computing environments. This is important because agents can run completely unsupervised risky because of their sometimes unpredictable nature.

With sandbox integration, agents can run in a stealth capacity within a specific workspace, accessing files and code only for specific operations, while otherwise protecting the overall integrity of the system.

Relatedly, the new version of the SDK also provides developers with a distributor attachment for boundary models, which will allow those agents to work with files and validated tools in the workspace, the company said. (In agent development, “harness” is a term that refers to components of an agent other than the model it runs on. An in-deployment harness often allows companies to both deploy and test agents running on boundary models. are considered the most advanced, general-purpose models available.)

“This launch is basically taking our existing agent SDK and making it compatible with all these sandbox providers,” Karan Sharma, who works on OpenAI’s product team, told TechCrunch.

The hope is that this, combined with the new harness capabilities, will allow users to “go build these long-horizon agents using our harnesses and using whatever infrastructure they have,” he said.

Such “long horizon” tasks are generally considered to be more complex and multi-stage work.

OpenAI said it will continue to expand the Agents SDK over time, but initially, the new harness and sandbox capabilities are being launched in Python first, with TypeScript support planned for a later release. The company said it is working to bring more agent capabilities, such as code mode and subagents, to both Python and TypeScript.

The new Agents SDK capabilities are offered to all customers via API and will use standard pricing.