[控场AI]
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/680013
API
curl https://kongchang.com/api/v1/knowledge/claims/680013
MCP
get_claim(id=680013)