Improve BLE reliability with notify retry and longer supervision timeout, add WiFi RSSI logging

Optimize face recognition: show boxes immediately, faster embeddings

Fix disconnect: stay disconnected when user taps Disconnect

Configure AVAudioSession for speech playback

Add text-to-speech: iPhone speaks recognized contact names

Add processing timing, skip auto-snapshots while recognition is running

Show raw snapshot immediately, then overlay annotations when ready

Add enrollment debug logging, skip contacts with no face in photo

Use full-size contact photos and lower match threshold to 0.25

Add debug logging to face recognition pipeline

Disable PicoClaw auto-upload for now

Enable CoreML model loading for face recognition

Add Windows CMD commands to README

Add README with setup instructions and face recognition enablement guide

Disable CoreML model loading until MobileFaceNet.mlpackage is added