Skip to main content
The Output node marks the end of your agent’s execution. When your agent reaches an output node, it will complete the workflow, shut down the browser, and return the final result. Every agent workflow should have at least one output node to properly conclude the execution.

​
Outcomes

Outcomes define the possible states your execution can end in. Think of them as exit codes or status labels that help you understand how and why your agent finished. Instead of being limited to just “success” or “failure”, you can define custom outcomes that match your specific use case.

​
Common Use Cases

  • Success/Failure States: success, failed, error
  • Business Logic States: item_purchased, out_of_stock, login_required
  • Conditional Results: approved, rejected, needs_review

​
Example Outcomes

For an e-commerce checkout agent, you might define outcomes like:
  • purchase_completed
  • payment_failed
  • item_unavailable
  • session_expired

​
Best Practices

  • Use clear, descriptive names (e.g., payment_completed instead of done)
  • Define all possible end states your agent might encounter
  • Keep outcome names consistent across similar workflows
  • Use snake_case for outcome names

​
Result Schema

You can configure a custom JSON Schema to define the structure of data returned by the output node. This is optional but useful when you need structured data in a specific format. Our schema format follows OpenAI’s Structured Outputs specification, which is a subset of JSON Schema. Your schemma result will be returned by the API at execution_result.result. See the full guidance and examples in Output & Status → Result Schema.

​
When to Use Custom Schemas

  • You need structured data in a specific format
  • You’re integrating with other systems that expect certain fields
  • You want to ensure consistent output structure across executions
  • You need to extract multiple pieces of information (prices, names, descriptions, etc.)

​
Example Schema

A schema for a product scraping agent: 
{
  "type": "object",
  "properties": {
    "product_name": {
      "type": "string",
      "description": "The name of the product"
    },
    "price": {
      "type": "number",
      "description": "The product price"
    },
    "in_stock": {
      "type": "boolean",
      "description": "Whether the product is available"
    }
  },
  "additionalProperties": false,
  "required": ["product_name", "price", "in_stock"]
}

​
Schema Requirements

When defining a schema, follow these rules:
  • All objects must include "additionalProperties": false
  • All object properties must be required - include all property keys in the required array
  • Use the description field to guide the AI in populating fields correctly

​
Tips

  • Start Simple: Begin with just outcomes, add a custom schema only if you need structured data
  • Test Your Outcomes: Make sure your agent logic can reach all defined outcomes
  • Validate Your Schema: Use the schema examples as a starting point to ensure proper formatting
  • Plan Your End States: Think through all the ways your agent’s task might conclude before defining outcomes