Circuit board: PCBA uses gold-immersion technology, the main control module uses ESP32-S3R8 chip, not a third party, good signal. More than 20 characters with different personalities and timbre can be configured; with memory, recent conversations can be summarized; adapted to multiple large models, including Deepseek. Wi-Fi/ML307 Cat.14G. BOOT key wake-up and interruption, support two triggering methods: click and long press. Offline voice wake-up ESP-SR. Streaming voice conversation (WebSocket or UDP protocol) Voiceprint recognition, identify who is calling the name of AI 3DSpeaker Large model TTS (Volcano Engine or CosyVoice) Large model LLM (Qwen, DeepSeek, Doubao) Configurable prompt words and timbre (custom role) Short-term memory, self-summary after each round of conversation Support LCD display of picture expressions Colour:Random Color(The color is random, and the designation is not accepted) Material:plastic Package Contents: 1 * AI Voice Chat Only the above package content, other products are not included. Note: Light reflection and different displays may cause the color of the item in the picture a little different from the real thing. The measurement allowed error is +/- 1-3cm.