Skip to main content

Some screenshots show mcp-s.com, which some existing organizations still use. New organizations use withwillow.ai. Copy the URLs shown in your own organization.

Background agents

A background agent is an AI agent you configure to do a particular job. You give it instructions and a set of tools, choose where it runs, and decide what starts its work. Its tool calls pass through the Willow gateway, where guards, policies, and logging apply.

For example, a repository briefing agent can read a GitHub repository and report its owner and default branch. A support agent might start when a ticket arrives, retrieve the relevant information, and prepare a response.

Open Build > Background Agents in the admin dashboard. This section is for administrators. If you use an agent assigned to you, open the end-user guide.

Build your first agent​

Create your first background agent walks through a small GitHub agent on Willow Agents: create it, add one read-only tool, connect the account, deploy it, and check its result.

Complete a task​

For other runtimes, use the platform reference to find the corresponding setup guide.

Understand how it works​

How background agents work explains instructions, tools, runtimes, authentication, and sessions. It also explains why an Active or Synced agent can still need tool authentication.

Look up settings and behavior​

Use the dashboard reference for list views, tabs, deployment states, and API behavior; the platform reference for runtime setup and Willow tokens; organization settings for owner permissions and available runtimes; and the credentials reference for gateway authentication and secret rotation.