Call an API from a flow

Use an API Call node for an HTTP integration you configure directly in the flow. You need the destination API’s URL, authentication method, accepted fields, and a test record.

Configure the request

Add an API Call node and choose GET, POST, PUT, PATCH, or DELETE.
Enter the URL. Add query parameters for values the API expects in the URL.
Add the required headers. Put the complete authentication value in Authorization, such as Bearer …, when the API uses that scheme.
For POST, PUT, or PATCH, choose the body format and supply the body. GET and DELETE in this node do not send a body.
Set a Response Variable, such as lookup_result.
Keep Wait for response enabled for this workflow and connect a Logic node to inspect the status before continuing.

{{variables}} can be used in the URL, headers, query values, and body. Raw JSON interpolation is text substitution: quotes or line breaks in a value can make your JSON invalid. Prefer simple validated values and test the exact request with the receiving service.

Read the result

With lookup_result as the response variable:

Variable	Contents
`lookup_result`	Response body as text, including JSON serialized as text
`lookup_result_status`	HTTP status as text, such as `200` or `404`
`lookup_result_error`	Network, timeout, or blocked-destination error when one is reported

An HTTP 400 or 500 response is still an HTTP response. Check the status and body; do not assume the error variable is populated for every unsuccessful status. Network failures use status 0. Requests have a bounded timeout, so design the receiving endpoint to respond promptly.

Verify before relying on it

Test success, a missing record, rejected credentials, and a slow or unavailable endpoint. For writes, inspect the destination record before retrying: a delayed response does not prove the write failed. If a destination is blocked, use a publicly reachable integration endpoint you control. Do not put secrets into visitor-facing messages or URL query strings. See Logic for status branches and Tool nodes for reusable actions.

API Call settings

The node picker calls this node API Call. Its request settings are:

Setting	Choices/default	Meaning and constraints
Name	`API Call` on a new node	Display label only.
Method	GET, POST, PUT, PATCH, DELETE; new nodes use POST	GET and DELETE send no body in this node.
URL	Empty initially	Destination URL; supports `{{name}}` substitution. Private/local destinations can be blocked.
Import from cURL	Paste a request	Populates request settings; review credentials and placeholders before saving.
Query Parameters	Key/value rows	Appended to the URL with URL encoding. Both keys and values support variables.
Authorization	Empty initially	Complete header value, such as `Bearer {{service_token}}`.
Headers	Key/value rows	Both keys and values support variables. A same-named configured header can replace Authorization; the selected body content type controls Content-Type when a body is sent.
Content type	JSON, form URL encoded, multipart form data, or plain text; default JSON	See body behavior below.
Body mode	Parameters or raw; default Parameters	Parameters represent key/value text; raw supplies the complete body.
Response Variable	Empty initially	Set a name to retain response text/status for later nodes. Empty means no named response fields are saved.
Wait for response	On by default	Keep on for a normal connected request. Turning it off currently skips the request on the main path; it does not create a reliable background request.

Body format	Current serialization
JSON + Parameters	JSON object whose keys and values are strings after substitution; `5` is sent as `"5"`.
JSON + raw	Raw substituted text. You must provide valid JSON and correct escaping.
Form URL encoded + Parameters	Encoded key/value pairs.
Form URL encoded + raw	Raw substituted text; supply correctly encoded form data yourself.
Plain text	Raw body text. Parameter rows do not generate a plain-text body.
Multipart form data	The editor offers the content type, but the node does not construct multipart boundaries or file parts. Use a supported JSON/form endpoint or a separately configured tool for uploads.

The HTTP request uses a 10-second timeout while awaiting a response; it is not a configurable node option or an end-to-end completion guarantee. A response error from a previous run can remain in <name>_error; always inspect the current <name>_status and current result. Async edges have separate execution behavior and are not a substitute for the Wait for response switch.

Start here

Build assistants

Connect channels

Manage & troubleshoot

Configure the request

Read the result

Verify before relying on it

API Call settings

​Configure the request

​Read the result

​Verify before relying on it

​API Call settings

Configure the request

Read the result

Verify before relying on it

API Call settings