Vercel
Vercel now supports building and deploying long-running, tool-using agents with the OpenAI Agents API. OpenAI manages the agent loop and session state. Vercel hosts the application and connects each session to Vercel Sandbox for code execution and file access.
According to the Vercel changelog, the integration uses signed OpenAI webhooks and Vercel Queues to create Sandboxes and reconnect agents to them reliably. Each agent session receives an isolated execution environment.
How OpenAI Agents API Runs on Vercel
The split gives each service a defined role. OpenAI controls the agent’s reasoning loop and keeps its session state. Vercel runs the surrounding application, manages the connection to Sandbox and provides the workspace where tools can execute code or work with files.
- 1
Start the session
OpenAI manages the agent loop and session state
- 2
Connect the Sandbox
signed webhooks and Vercel Queues handle creation and reconnection
- 3
Run the tools
the agent executes code and accesses files inside an isolated environment
- 4
Continue the task
follow-up instructions use the same persistent workspace
Files remain available across follow-up instructions. An agent can create or change a file, receive another request and continue working in the same workspace rather than starting again without that material.
The architecture also scales to zero. It does not require an always-on worker to remain active between jobs. Separate execution environments keep each agent session isolated from other sessions.
Building With the OpenAI Agents API
Vercel has published a step-by-step guide covering how to build and deploy an agent, alongside a sample application. This update also follows Vercel’s earlier support for Cursor Cloud Agents through Sandbox, covered in our report on that integration.
The new setup is designed for agents that need more than one short model response. It supports ongoing sessions, tool calls, code execution, file access and later instructions without requiring developers to operate the agent loop themselves.
Frequently asked questions
What is the OpenAI Agents API integration on Vercel?
It is a way to deploy long-running, tool-using agents while OpenAI manages the agent loop and session state. Vercel hosts the application and attaches every session to an isolated Sandbox for code execution and file access.
Is the OpenAI Agents API integration free?
The announcement does not include pricing or plan details. It confirms that a build guide and sample application are available.
When is the OpenAI Agents API integration available?
Vercel announced the integration on 10 September 2026 and described it as available now. Developers can follow the published guide to build and deploy an agent.
Sources
3 checkedHow we cover tool news: Create With's tool desk drafts these reports with AI from the sources listed above and checks them against those sources before publishing.
Supabase
