Unverified50% confidenceReleaseExact time
Microsoft has previously released the MAI-Voice-1 voice model and the MAI-1 foundation language model.
1
Sources
50%
Confidence
Long-term
Relevance
9/4/2026
First Seen
Sources
Related Entities
Related Claims
Unverified自回归神经语音模型的代表包括微软的VALL-E和VALL-E 270% similarUnverifiedThe software supports AI voice cloning that converts scripts to voice versions in the user's own voice with one click68% similarUnverifiedWindows' built-in voice input feature is powered by Microsoft Azure Cognitive Services' speech recognition engine65% similarVerified传统语音AI普遍采用级联式架构，先用ASR将语音转文字，再交给语言模型处理，最后用TTS合成输出65% similarUnverifiedSpeech synthesis models such as VALL-E and VoiceCraft adopt the technical approach of representing speech as token sequences and using an autoregressive generation framework.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/856886API
curl https://kongchang.com/api/v1/knowledge/claims/856886MCP
get_claim(id=856886)