待验证90% 置信事实时间未知
End-to-end latency in voice AI typically involves three stages: speech-to-text (STT) recognition, large language model inference, and text-to-speech (TTS) synthesis
1
来源数
90%
置信度
长期有效
时效性
2026/8/2
首次发现
来源
涉及实体
相关事实
引用此条事实
Stable URI
https://kongchang.com/claim/680013API
curl https://kongchang.com/api/v1/knowledge/claims/680013MCP
get_claim(id=680013)