Analytics preference

We only load Google Analytics after you opt in. You can change this choice at any time from Cookie settings.

SSE Streaming

Workflows can stream execution events in real time with Server-Sent Events (SSE).

Workflows can stream execution events in real time with Server-Sent Events (SSE).

Endpoint

POST /api/workflows/{workflow_id}/execute/stream

This endpoint uses the same workflow-level authentication, rate limiting, and response caching rules as the standard execute endpoint documented in Webhooks.

Enabling SSE in the cURL Dialog

Open the Run with cURL dialog in the editor and enable SSE Streaming.

When enabled, Heym:

  • switches the generated command from /execute to /execute/stream
  • adds --no-buffer so streamed events appear immediately in terminal output
  • adds Accept: text/event-stream
  • lets you configure per-node node_start messages directly in the dialog

Event Protocol

Each SSE frame is newline-delimited JSON prefixed with data: .

Event typeWhen it is emittedKey fields
execution_startedImmediately after the request is acceptedexecution_id
node_startRight before a node begins executionnode_id, node_label, optional message
node_retryWhen a node retriesnode_id, attempt, max_attempts
node_completeAfter a node finishesnode_id, node_label, status, output, execution_time_ms
final_outputWhen an Output node returns earlynode_label, output
execution_completeAfter the workflow endsstatus, outputs, execution_time_ms

Per-Node Start Messages

The cURL dialog stores SSE settings on the workflow itself:

  • Send start controls whether a node_start event includes a message
  • Start message defaults to [START] {nodeLabel}
  • clearing a custom message falls back to the default text
  • node_complete is always emitted and always includes the node output

Example stored config:

{
  "node-uuid-llm": {
    "send_start": true,
    "start_message": "[START] Draft reply"
  },
  "node-uuid-http": {
    "send_start": false,
    "start_message": null
  }
}

cURL Example

curl -X POST --no-buffer \
  -H "Content-Type: application/json" \
  -H "Accept: text/event-stream" \
  -H "Authorization: Bearer <your-access-token>" \
  "https://app.heym.ai/api/workflows/{workflow_id}/execute/stream" \
  -d '{"message":"Hello"}'

Sample event stream:

data: {"type":"execution_started","execution_id":"..."}
 
data: {"type":"node_start","node_id":"...","node_label":"LLM","message":"[START] LLM"}
 
data: {"type":"node_complete","node_id":"...","node_label":"LLM","status":"success","output":{"text":"Hello world"},"execution_time_ms":312}
 
data: {"type":"execution_complete","status":"success","outputs":{"LLM":{"text":"Hello world"}},"execution_time_ms":315}

Security Notes

  • anonymous, jwt, and header_auth workflows behave the same as the regular webhook endpoint
  • response caching still applies when cache_ttl_seconds is configured
  • workflow rate limits still apply when rate_limit_requests and rate_limit_window_seconds are configured
  • test_run bypasses both cache and rate limiting, the same way it does for /execute