PROTOCOLS

Server-Sent Events

Real-time streaming over HTTP — auto-detected, with a live event viewer, mid-stream cancellation, and full history persistence.

Auto-Detection

SSE is not a separate protocol mode — it works with regular HTTP requests. No protocol selector change needed.

Auto-detected when the response Content-Type is text/event-stream

The Send button changes to Stop while streaming

Can also be forced via the isSSE flag on a saved request

Works with any HTTP method — GET or POST (e.g. OpenAI streaming uses POST)

Protocol selector stays on HTTP — no change needed

Key point: Unlike WebSocket or gRPC which require a protocol switch, SSE is just a regular HTTP response with a streaming body. You build your request exactly like any other HTTP request.

Stream Tab

A Stream tab appears in the response viewer when SSE is detected. Events render live as they arrive — no waiting for the stream to end.

Each event shows: event type (if specified), data payload, timestamp, and optional id

Events are numbered sequentially (1, 2, 3…)

Total event count and stream duration shown in the tab header

The Body tab shows the accumulated raw text (all events concatenated)

JSON pretty-printing: If an event's data payload is valid JSON, it is automatically pretty-printed in the Stream tab for readability.

Event Format

API Studio supports the standard SSE wire format:

event: message
data: {"text": "Hello"}
id: 1

event: message
data: {"text": "World"}
id: 2

Field	Description
event:	Optional event type name (defaults to "message")
data:	The payload — can be multi-line (multiple `data:` lines are joined)
id:	Optional event ID for reconnection tracking
\n\n	Double newline separates events

Multi-line data: fields are automatically joined into a single payload in the Stream tab display.

Cancelling Mid-Stream

Click the Stop button to cancel the stream at any time.

Events received so far are preserved in the Stream tab

Response time shows as stream duration (total time connected)

History entry saves all events received up to cancellation

Common Use Cases

OpenAI / Anthropic chat completions with "stream": true

Real-time notifications and feeds

Log streaming and monitoring

CI/CD build status updates

Examples

OpenAI streaming

POST https://api.openai.com/v1/chat/completions

{
  "model": "gpt-4",
  "messages": [{"role": "user", "content": "Hello"}],
  "stream": true
}

Public test endpoint

GET https://sse.dev/test

Tips

SSE responses can be large — the Body tab accumulates all data

For very long streams, consider cancelling periodically to free memory

SSE works through proxies that support chunked transfer encoding

If auto-detection doesn't trigger, set isSSE: true on the saved request

Auth headers work normally — SSE is just HTTP with a streaming response

History saves an events[] array with id, event, data, and timestamp per event

GraphQL → Sending Requests →