Unify Logo Footer.svg
Unify Automations
Logo
Execute Streaming REST request

Execute Streaming REST request

Logo

2 mins READ

Execute Streaming REST Request sends an HTTP request and processes the response as a stream of chunks rather than waiting for the full body to arrive. Use it for server-sent events, chunked transfer responses, and any API that delivers data progressively — such as LLM inference endpoints or real-time data feeds.

Overview

Some APIs send their responses as a continuous stream of chunks rather than a single complete payload. The Execute Streaming REST Request action keeps the HTTP connection open and yields response chunks as they arrive, allowing your automation to act on partial data before the stream ends. The connection closes when the server finishes sending or when an error interrupts the stream.

Run Asynchronously is not available for streaming requests — the stream is consumed progressively by design, and the automation processes each chunk as it arrives rather than waiting for a single final response.

Screenshot_2026-08-30_at_12.19.13_AM_1.png
Screenshot_2026-08-30_at_12.19.13_AM_1.png

Input

Field

Description

Required*

Method

HTTP method for the request: GET or POST. Streaming APIs most commonly use GET for Server-Sent Events and POST for chunked inference or data APIs.

Yes

URL

The streaming endpoint URL.

Yes

Headers

Request headers. For Server-Sent Events, include Accept: text/event-stream. For bearer-token APIs, include Authorization: Bearer <token>.

No

Query Parameters

Key-value pairs appended to the URL as a query string. Values are URL-encoded automatically.

No

Request Body

Body payload to send with a POST request (e.g., a JSON object with "stream": true for an LLM API).

No

Authentication Type

Authentication scheme: NoneBasic, or Custom.

No

Username

Username for Basic authentication. Active only when Authentication Type is Basic.

No

Password

Password for Basic authentication. Active only when Authentication Type is Basic.

No

Custom Auth Configuration

Configuration object for a custom authentication scheme. Active only when Authentication Type is Custom.

No

SSL Verification

Verify the server's TLS certificate. Enabled by default; disable only in trusted, non-production environments.

No

Proxy URL

Route the request through an HTTP or HTTPS proxy. Note that some proxies buffer responses, which can interfere with streaming.

No

* Required fields must be populated for the node to run.

Output

Field

Description

Status Code

The HTTP status code returned when the server accepts the streaming connection (e.g., 200).

Response Stream

The stream of chunks yielded progressively as the server sends data. Each chunk is a raw string; parse individual chunks as needed.

Response Headers

A map of the HTTP response header names and values received when the connection was established.

Notes

  • Run Asynchronously is not available for streaming requests. The stream is consumed progressively — the automation processes chunks as they arrive rather than collecting a final response.

  • For Server-Sent Events (SSE), set the Accept header to text/event-stream. The node reads the stream line by line according to the SSE format.

  • A non-streaming API called through this action will return its full body as a single chunk. This is valid but offers no advantage over Execute REST Request for non-streaming endpoints.

  • HTTP 4xx and 5xx status codes are reported in the Output Status Code field. The node does not raise them as errors — handle them in subsequent steps.

  • Some proxies buffer entire responses before forwarding them. If streaming does not work as expected, try a direct connection by removing Proxy URL or switching to a streaming-compatible proxy.

FAQs

Does this action support HTTP/2 server push?

The node works over HTTP/1.1 chunked transfer and HTTP/2 streaming, depending on what the server and your network infrastructure support. HTTP/2 server push is a separate mechanism and is not directly controlled through this node's configuration.

What happens if the stream connection drops unexpectedly?

The node ends the step with whatever chunks were received before the connection dropped. Implement retry logic in your automation if guaranteed delivery is required.

Can I use this node to call an LLM streaming API?

Yes. Set Method to POST, URL to the inference endpoint, add your API key in Headers as Authorization: Bearer <key>, and set the Request Body to your JSON payload with "stream": true. Response chunks arrive as the model generates tokens.