StreamClipper AI Faz 1: ingestion + sinyal analizi + STT + altyapı iskeleti

yt-dlp headless capture (15dk segmentleme), ses peak + chat velocity sinyal
tespiti, OpenAI Whisper STT (kelime zaman damgalı), BullMQ/Redis/Postgres
altyapısı ve kanal durumu + transkript kütüphanesi gösteren Next.js panel.
LLM virality skorlama, render/crop/altyazı ve multi-platform dağıtım bu
fazın kapsamı dışında.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

apps/api-daemon/src/queues.ts

@@ -0,0 +1,36 @@
+import { Queue } from "bullmq";
+import { redisConnection } from "./redis";
+
+/**
+ * Queue names are shared across the Node (api-daemon) and Python (worker)
+ * services via the protocol-compatible `bullmq` package on both sides.
+ * Keep this list in sync with apps/worker/worker/queues.py.
+ */
+export const QUEUE_NAMES = {
+  STREAM_INGEST: "stream-ingest",
+  SIGNAL_DETECTION: "signal-detection",
+  STT_SCORING: "stt-scoring",
+  // Defined for future modules (render/distribution) — no processor yet.
+  VIDEO_RENDER: "video-render",
+  POST_PUBLISH: "post-publish",
+} as const;
+
+export interface StreamIngestJob {
+  sessionId: string;
+  channelDbId: string;
+  youtubeUrl: string;
+}
+
+export interface SignalDetectionJob {
+  rawSegmentId: string;
+  filePath: string;
+  sessionId: string;
+}
+
+export const streamIngestQueue = new Queue<StreamIngestJob>(QUEUE_NAMES.STREAM_INGEST, {
+  connection: redisConnection,
+});
+
+export const signalDetectionQueue = new Queue<SignalDetectionJob>(QUEUE_NAMES.SIGNAL_DETECTION, {
+  connection: redisConnection,
+});