Auto Caption

Live caption generation engine with provider selection and WebSocket streaming sessions.

프로바이더 선택과 WebSocket 스트리밍 세션을 지원하는 라이브 자동 자막 엔진입니다.

Overview

Contract Scope

The live-caption engine is running in production under /api/live-captions/*.

The capabilities endpoint reports configured:true, status:active, and providers [mock, openai, rtzr, kangkyu_beta].

Caption sessions stream updates over WebSocket at /live-captions/{sessionId}/stream.

API Reference

Endpoints

GET

/api/live-captions/capabilities

Read live caption engine capability and provider state.

Response

{ configured:true, status:'active', providers:['mock','openai','rtzr','kangkyu_beta'] }.

This endpoint is live in thegreit.com production.

POST

/api/live-captions/sessions

Create a caption session with an optional provider choice.

Name	Type	Required	Description
provider	mock \| openai \| rtzr \| kangkyu_beta	No	Provider selected for the caption session.
audio	file \| stream reference	No	Audio source used by the live caption session.

Response

Session metadata, including the session id used by the stream route.

Provider availability follows the capabilities response.

/live-captions/{sessionId}/stream

Stream caption events for an active session.

Name	Type	Required	Description
sessionId	string	Yes	Caption session identifier returned by the session creation flow.

Response

Streaming caption events with timing and text updates.

WebSocket streaming is part of the live production engine.

Quickstart

Five-line curl

curl

curl "https://thegreit.com/api/live-captions/capabilities" \
  -H "Authorization: Bearer $GREIT_SESSION_TOKEN" \
  -H "Accept: application/json" \
  -H "X-Client: docs-quickstart" \
  -i

Version log

Release Notes

current

2026-07-07

Live production engine documented with capabilities, provider selection, and WebSocket streaming.
Providers documented from audit: mock, openai, rtzr, kangkyu_beta.