init: Kira — AI body double with Honcho memory

Full voice pipeline (Whisper STT -> DeepSeek LLM -> OpenAI TTS),
animated SVG avatar (Live2D-ready), girly-pop UI, lofi music,
timer/notes/pets/wardrobe widgets, 10 background scenes with
particle effects, Honcho cross-session memory.

docker-compose.yml

@@ -0,0 +1,27 @@
+services:
+  backend:
+    build: ./backend
+    container_name: kira-backend
+    env_file: .env
+    volumes:
+      - ./.env:/app/.env:ro
+    ports:
+      - "8000:8000"
+    restart: unless-stopped
+    networks:
+      - kira-net
+
+  frontend:
+    build: ./frontend
+    container_name: kira-frontend
+    ports:
+      - "3000:80"
+    depends_on:
+      - backend
+    restart: unless-stopped
+    networks:
+      - kira-net
+
+networks:
+  kira-net:
+    driver: bridge