docs: add emoji gesture reference table

v0.6.2: Gesture detection with HaGRID ONNX models (18 gestures)

Fix: push gesture state to Home Assistant via callback

Debug: add hand detector raw output logging, lower thresholds to 0.3

debug: 添加手势检测循环确认日志

debug: 添加手势检测INFO日志

fix: 重写gesture_detector.py，修正模型输出格式

fix: 重写手势检测器，使用正确的模型输入输出格式

fix: 使用状态判断唤醒词检测, 添加手势模型信息日志

fix: 修复DOA方向, 减少日志, 添加手势检测调试

fix: 修复DOA方向，减少语音日志输出

fix: 修复手势检测结果不更新的问题

chore: 添加手势识别ONNX模型文件 (LFS)

feat: 添加手势识别功能 (HaGRID ONNX模型)

remove duplicated wakeword

update model name

docs: change wake word example from Hey Jarvis to Hey Reeche

docs: highlight Multi-room audio (Sendspin) in hero badges

docs: update index.html - remove continuous conversation highlight, update entity count to 45+

docs: complete changelog history from v0.1.0 to v0.6.1

refactor: load changelog from JSON file dynamically

chore: bump version to 0.6.1, show only 6 changelog entries by default

feat: prioritize MicroWakeWord over OpenWakeWord for same-name wake words

fix: 加载openWakeWord子目录中的唤醒词配置

Upload hey_ree_che.tflite

fix: 修复stop词检测，设置stop_word的id属性

feat: 添加 hey_ree_che 唤醒词配置

fix: 使用固定大小音频块发送，提升STT响应速度

remove unused file

release: v0.6.0 - 实时音频驱动说话动画

feat: 添加实时音频驱动的说话动画

fix: 移除音频预下载分析，使用JSON动画替代

refactor: 移除硬编码动作，动画完全由JSON配置驱动

fix: 调整思考动画，减少向后仰头幅度

feat: 添加基于音频分析的说话动画

improve: 改进说话动画，使用多频率振荡器模拟自然头部摆动

update index.html link

improve: 增强说话动画幅度，更自然地模拟人类讲话

debug: 将 Voice event 日志改为 INFO 级别

debug: 添加说话动画的调试日志

refactor: 简化唤醒词检测逻辑，按照参考项目模式

Fix: Restore HA continue_conversation support, our switch is unconditional override

Fix: Ignore HA continue_conversation, only use our switch

v0.5.16: Remove tap-to-wake, add Continuous Conversation switch

Fix: PEP8 compliance and wake word refractory period after conversation ends

Fix: asyncio.coroutine removed in Python 3.12, use Coroutine type hint

v0.5.15: Audio settings persistence, Sendspin discovery refactor, tap detection fix

CRITICAL FIX: Rename duplicate _reachy_on_listening to _reachy_on_idle - tap detection was never re-enabled

Speaking: pause face tracking, use lively nodding animation (1.5Hz)

Add logging to debug tap detection disable