Skip to main content

Build Custom Functions with the Code Component

The code component allows you to execute product or industry-specific code within an integration. This page outlines when and how to use a code component.

Why use the code component?

You will likely have integration logic that can't be solved using the standard components Prismatic provides. The portion of your integrations that are specific to your product or industry can be accomplished using code component steps, or custom components.

When to build a custom component instead

Code component steps should be succinct and integration-specific. If the code you write could be reused in other integrations, or if the code is complex enough that it would benefit from unit tests, etc., you should build a custom component instead.

Adding a code step to an integration

Within the integration designer, add a step to your integration. Select the Code component, Code Block action.

You will be presented with a new code step in your integration, and you can click the "Edit" button to open the code editor:

Code structure

The code component provides a stub function by default. Let's examine the structure of the code:

module.exports = async ({ logger, configVars }, stepResults) => {
return { data: null };

The code component requires you to export an asynchronous function. The default code uses arrow function notation to create an async function to return.

Code component parameters

This function is provided a few parameters:

  1. The first positional parameter, is comprised of several properties:
    • logger allows you to write out log lines.
    • configVars lets you access config variables (including connections).
    • instanceState, crossFlowState, integrationState and executionState gives you access to persisted state.
    • stepId is the ID of the current step being executed.
    • executionId is the ID of the current execution.
    • webhookUrls contains the URLs of the running instance's sibling flows.
    • webhookApiKeys contains the API keys of the running instance's sibling flows.
    • invokeUrl was the URL used to invoke the integration.
    • customer is an object containing an id, name, and externalId of the customer the instance is assigned to.
    • user is an object containing an id, name, email (their ID) and externalId of the customer user whose user-level config was used for this execution. This only applies to instances with User Level Configuration.
    • integration is an object containing an id, name, and versionSequenceId of the integration the instance was created from.
    • instance is an object containing an id and name of the running instance.
    • flow is an object containing the id and name of the running flow.
  2. The second positional parameter, stepResults, is an object that contains output from previous steps of the integration.


context.logger is an object that can be used for logging and debugging. context.logger has four functions: debug, info, warn, and error. For example:

module.exports = async (context, stepResults) => {"Things are going great");
context.logger.warn("Now less great...");

// or

module.exports = async ({ logger }, stepResults) => {"Hello World");

Note: Log lines are truncated after 4096 characters. If you need longer log lines, consider streaming logs to an external log service.

Config variables

context.configVars provides the Code Component with access to all config variables, including connections, associated with the integration.

If you have a config variable named Acme ERP Base URL, for example, you could reference that config variable in a code step with context.configVars["Config Variable Name"] syntax:

module.exports = async ({ configVars }, stepResults) => {
const fuelApiUrl = `${configVars["Acme ERP Base URL"]}/fuel`;
// ...


Connections are a special type of config variable. You can access the contents of a connection the same way that you would any other config variable. In this example, suppose you have a connection config variable named Acme ERP Connection that contains two fields, tenantId and apiKey:

Destructuring a connection config variable
module.exports = async ({ logger, configVars }, stepResults) => {
const {
fields: { tenantId, apiKey },
} = configVars["Acme ERP Connection"];
const result = await doAThing({ tenantId, apiKey });
return { data: result };

Referencing previous step outputs

Most steps of an integration return some sort of value. An HTTP - GET action, for example, might return a JSON payload from a REST API. An Amazon S3 - Get Object will return a binary file pulled from S3.

The code component can reference those outputs through the stepResults parameter. stepResults is an object that contains results from all previous steps.

For example, if you have an HTTP - GET step named Fetch Users List that pulls down an array of users from, you can generate an array of email addresses with this code:

module.exports = async (context, stepResults) => {
const userArray = stepResults.fetchUsersList.results;
const emailArray = =>;
return { data: emailArray };
Step results are often objects

Many components return objects that have multiple keys. So, you can reference stepResults.myStepName.results.someKey. It's rare for a component to return serialized JSON, so there's rarely need to JSON.parse() results from a previous step.

Previous steps names as variables

Since names of steps can include spaces and non-JavaScript-friendly characters, alphanumeric characters of step names are converted to camelCase. So, a step named Download JSON from API would be converted to downloadJsonFromApi. You can test out step name -> referenceable name conversions here:

Referencing integration trigger payload data

The integration trigger is simply another step that can have a unique name. Suppose an integration is triggered by a webhook, the trigger is named My Integration Trigger, and the webhook is provided a payload of {"exampleKey": "exampleValue"}.

That exampleKey would be accessible using stepResults.myIntegrationTrigger like so:

module.exports = async ({ logger }, stepResults) => {
const exampleKey =;`Received '${exampleKey}'`);

Using JavaScript destructuring, you can instead write this:

module.exports = async (
{ logger },
myIntegrationTrigger: {
results: {
body: {
data: { exampleKey },
) => {`Received '${exampleKey}'`);

Notice the logged message in the testing drawer

Persisted data in a code step

Like a custom component, a code step can save and reference persisted state at the flow (instanceState), cross-flow (crossFlowState), integration (integrationState) or execution (executionState) level.

To save a value bar as key foo at the executionState level:

module.exports = async ({ logger, configVars }, stepResults) => {
return {
data: null,
executionState: { foo: "bar" },

To load the value of the key foo at the execution level you can reference your function's first parameter's executionState property:

module.exports = async ({ executionState }, stepResults) => {
const myvalue = executionState["foo"];
return { data: `My value is ${myvalue}` };

Code component return values

The code component can optionally return a value for use by a subsequent step. The return value can be an object, string, integer, etc., and will retain its type as the value is passed to the next step.

The return value is specified using the data key in the return object.

In this example we return a string with value "":

module.exports = async (context, stepResults) => {
return { data: "" };

The output can be used as input for the next step by referencing codeComponentStepName.results.

To see an example of returning more complex data structures, and a good example use case for a code component, see the Using a Code Component to Transform Data quickstart.

Returning binary data from a code step

Sometimes you'll want your code component to return binary data (like a rendered image or PDF). To do that, return an object with a data property of type Buffer (a file buffer), and a contentType property of type String that contains the file's MIME type:

module.exports = async (context, stepResults) => {
// ...
const fileBuffer = SomePdfLibrary.generatePdf();
return {
data: fileBuffer,
contentType: "application/pdf",

To see an example use case for returning binary data from a code component, check out our Generate a PDF with a Code Component.

For More Information: Using a Code Component to Transform Data

Making HTTP calls from a code step

The fetch API is baked into the code component. To make an HTTP call to an API, you can use the fetch function:

Make an HTTP POST request from a code step
module.exports = async (context, stepResults) => {
const options = {
method: "POST",
headers: {
Accept: "application/json",
"Content-Type": "application/json",
Authorization: "Bearer abc-123",
body: JSON.stringify({ foo: "bar", baz: 123 }),
const response = await fetch("", options);
const jsonData = await response.json();
return { data: jsonData };

Adding dependencies to a code step

If your code component depends on node modules from npm, dependencies will be dynamically imported from the UNPKG CDN. For example, if your code component reads:

Import lodash as a dependency
const lodash = require("lodash@4.17.21/lodash.js");

module.exports = async (context, stepResults) => {
const mergedData = lodash.merge(
{ cpp: "12" },
{ cpp: "23" },
{ java: "23" },
{ python: "35" }
return { data: mergedData };

Then lodash version 4.17.21 will be imported as a dependency.

You should specify specific known working versions of npm packages for your code component:

const lodash = require("lodash@2.4.2");
const { PDFDocument } = require("pdf-lib@1.17.1/dist/pdf-lib.js");

You can require any file from npm using package[@version][/file] syntax. Note that with the lodash import above, no file was specified. If no file is specified, the main file defined in the npm package's package.json is imported. An explicit path was called out for the pdf-lib import because the pdf-lib package defaults to importing an index file that itself requires other files, and dist/pdf-lib.js is a completely independent file that can be imported on its own..

CDN outages and downstream dependencies

Dynamic imports may or may not work, depending on how the dependency maintainer compiled their package. All JavaScript code must be compiled into a single file to be dynamically imported. If the package has its own dependencies that are not compiled in, or if the file you reference has its own require() statements, you may see errors in your code.

You might also see errors if the UNPKG CDN is unavailable for any amount of time.

If you need external dependencies, we strongly recommend using a code step for prototyping, but building a custom component for production use. Custom components have their dependencies compiled in, and are not dependent on the uptime of an external CDN.

Requiring built-in NodeJS modules

You can also require built-in NodeJS modules, like crypto or path. If the library you specify is built in to NodeJS, the client will not reach out to the UNPKG CDN, and will instead use the built-in module.

const crypto = require("crypto");

module.exports = async () => {
const { publicKey, privateKey } = crypto.generateKeyPairSync("rsa", {
modulusLength: 4096,
publicKeyEncoding: {
type: "spki",
format: "pem",
privateKeyEncoding: {
type: "pkcs8",
format: "pem",
cipher: "aes-256-cbc",
passphrase: "top secret",

return {
data: {

Using spectral utility functions in a code step

Prismatic's SDK, @prismatic-io/spectral is automatically imported into each code block as spectral. You can reference any utility functions that Spectral declares.

For example, if you need to cast a truthy "NO" string to a boolean, you can do this:

module.exports = async ({ configVars }, stepResults) => {
const doAThing = spectral.util.types.toBool(configVars["Do a Thing?"]);
if (doAThing) {
return "Did a thing";
} else {
return "Didn't do a thing";

A list of all util type functions is available in the Spectral SDK docs.