Webhooks and AI agents: trigger any system from a workflow

What you can build

How to connect Webhooks

1. 01Create a tool of type webhook with the target URL, an event name and any headers the receiver expects, such as a shared-secret header.
2. 02Describe the input schema so the model knows which fields to include in the payload.
3. 03Set the safety level write_safe, a timeout, a retry policy and rate limits; mark it as requiring approval if a human should confirm each trigger.
4. 04Attach the tool to the agent, fire it from the test chat and confirm the payload in the run trace and in the receiving system.

Example tool definition

The shape a tool takes in the AgentRow builder API. Credentials are stored on the tool's auth settings, not in its configuration, so they never reach a prompt. Placeholders in double braces are filled by the model from the input schema.

{
  "name": "notify_lead_pipeline",
  "type": "webhook",
  "description": "Send a qualified lead to the sales pipeline. Use once the user has shared a name, company and email and asked to be contacted.",
  "auth_type": "custom_header",
  "safety_level": "write_safe",
  "config": {
    "url": "https://hooks.example.com/agentrow/leads",
    "event_name": "lead.qualified"
  },
  "input_schema": {
    "type": "object",
    "properties": {
      "name": { "type": "string" },
      "company": { "type": "string" },
      "email": { "type": "string" },
      "summary": { "type": "string", "description": "One-paragraph summary of what the lead needs" }
    },
    "required": ["name", "email"]
  }
}

What the run trace shows

Every call appears in the run trace with the rendered request, the response status and body, timing, retries and errors, plus what the model was shown of the result. If the model never called the tool, the toolbox step shows whether it was bound for that model call and why.