# Core dependencies
numpy>=1.24.0
opencv-python>=4.8.0
sounddevice>=0.4.6

# ML models
open-clip-torch>=2.23.0  # CLIP ViT-B/32 for visual semantics
faster-whisper>=0.10.0   # Whisper ASR for transcription
sentence-transformers>=2.2.0  # MiniLM-L6-v2 for text embedding

# Phase 2: WebSocket server
fastapi>=0.109.0
uvicorn[standard]>=0.27.0
websockets>=12.0

# Utilities
pillow>=10.0.0  # Image processing for CLIP

# Development (optional)
pytest>=7.4.0
# scipy>=1.11.0  # For weight calibration
