Skip to main content

​
Overview

The Code step lets you run custom scripts with full programming flexibility. Use it when you need logic that goes beyond what the Transformer or Validator can do — custom calculations, calling external APIs with complex auth flows, parsing files, or any operation you’d write in a script. Scripts run in an isolated sandbox with a 60-second timeout. Each execution gets a clean environment, so there is no shared state between runs.

​
Runtimes

RuntimeVersionAsync Support
Node.jsv22Top-level await works out of the box (scripts run inside an async wrapper)
Python3.13Standard async/await with asyncio

​
The input Object

Your script has access to a global input object that contains the full execution context:
  • Trigger data — the event or manual payload that started the workflow
  • Previous step outputs — results from all steps that ran before this one
  • Database records — data from connected automation databases
// Node.js — access previous step output
const apiResponse = input.steps["fetch-data"].output.body;
const contractId = input.trigger.contract_id;
respond({ contractId, data: apiResponse });

# Python — access previous step output
api_response = input["steps"]["fetch-data"]["output"]["body"]
contract_id = input["trigger"]["contract_id"]
respond({"contract_id": contract_id, "data": api_response})

​
Configuration

FieldDescription
RuntimeNode.js or Python
ScriptYour code in the built-in editor
AuthOptional JSON credentials secret for the script to use

​
Returning Output

Your script must call respond(data) to return output to the workflow. If the script completes without calling respond(), the step will fail. 
// Node.js — return a result
const result = { processed: true, count: 42 };
respond(result);

# Python — return a result
result = {"processed": True, "count": 42}
respond(result)
To signal an error, pass a second argument: 
respond(null, "Something went wrong");

​
JSON Credentials

When you select JSON Credentials as the auth type and attach a secret, the credential file is written into the sandbox. Access the file path via the CREDENTIALS_PATH environment variable: 
// Node.js — use JSON credentials
import { readFileSync } from "fs";

const creds = JSON.parse(readFileSync(process.env.CREDENTIALS_PATH, "utf8"));
// Use creds to authenticate with an external service

# Python — use JSON credentials
import json, os

with open(os.environ["CREDENTIALS_PATH"]) as f:
    creds = json.load(f)
# Use creds to authenticate with an external service

​
Editor Features

The script editor includes:
  • Expand — open the editor in a fullscreen dialog for more space
  • Generate with AI — click the sparkle icon to have AI generate a script from a natural language description
  • Syntax highlighting — JavaScript or Python highlighting based on the selected runtime

​
Using Libraries

The script editor has built-in library integration. Click Library above the editor to load a saved script, or click Save to store the current script as a reusable library item. This is useful when the same logic is shared across multiple automations.

​
Output

FieldTypeDescription
dataanyThe data passed to respond()
successbooleanWhether the script completed successfully
errorstringError message if the script failed