Overview

Code tasks support direct placeholders inside the code editor. You no longer need separate "input field mapping" for most code-task scenarios. Add placeholders where values are needed and FlowFn resolves them at runtime.

Placeholder syntax

{{$trigger.<fieldKey>}} – value from trigger inputs.
{{$task.<taskId>.outputs.<fieldKey>}} – value from a previous task output.

Examples:

const email = {{$trigger.email}};
const customerId = {{$task.fetchCustomer.outputs.customer_id}};
const amount = {{$task.calculate.outputs.total_amount}};

Using custom values in code

You can mix literal constants and placeholders in your code:

const payload = {
  team: "sales",
  threshold: 100,
  userEmail: {{$trigger.email}},
  summary: {{$task.summarize.outputs.text}}
};

For string interpolation, keep placeholders in JavaScript expression positions (not inside quoted strings):

const message = `Hello ${{$trigger.first_name}}!`;

Type behavior

FlowFn resolves placeholders with type-aware casting so values are preserved when possible (number, boolean, date/datetime/time, json/model, array). File fields resolve to the full file object (artifact pointer or URL metadata) unless you use a dotted subpath: .url is a presigned read URL when the file is in S3 (for example {{$task.<id>.outputs.my_file.url}}); other common fields include .key, .bucket, .region, .filename (artifact metadata), and for inline buffers .contentType or .name, depending on how the file was produced.

Best practices

Reference outputs only from tasks that run earlier in execution order.
Use clear output keys in upstream tasks so placeholders remain readable.
Test with a manual run after changing placeholders.
If a value may be missing, guard it in code (for example with default fallback logic).

Troubleshooting

If a placeholder resolves empty, check the task ID/key and verify the upstream task produced that output.
If runtime behavior looks different than expected, inspect run details and output payloads for each step.