xAI opens Grok STT and TTS audio APIs, with overall word error rate of STT reduced to 6.9%

robot
Abstract generation in progress
ME News Report, April 18 (UTC+8), according to Beating Monitoring, xAI has launched two standalone audio APIs: Grok Speech to Text and Grok Text to Speech. Both come from the same audio stack supporting Grok Voice, Tesla's in-car system, and Starlink customer service, now available as independent endpoints, allowing developers to directly connect to voice agents, real-time transcription, accessibility tools, and podcasts. STT offers two modes. The REST API is used for batch transcription of large audio files with millisecond-level response; the WebSocket API is designed for real-time audio streams. Features include word-level timestamps, speaker diarization, multi-channel recognition, and Inverse Text Normalization, which automatically formats spoken numbers, dates, and currencies into standardized structured text. Supported languages exceed 25, with seamless switching during conversations. xAI also released a set of Word Error Rate (WER) comparisons: overall scene Grok 6.9%, ElevenLabs 9.0%, Deepgram 11.0%, AssemblyAI 12.9%; the gap in "Phone Call Entity Recognition" is even larger, with Grok at 5.0%, compared to 12.0%, 13.5%, and 21.3% for the other three. In common business scenarios like meetings, video podcasts, and phone calls, Grok also maintains a slight lead. These figures are self-tested and published by xAI, with no third-party verification yet. In terms of pricing, batch STT is $0.10 per hour, streaming is $0.20 per hour; TTS costs $4.20 per million characters. TTS supports inline Speech Tags to control emotion and intonation, such as [laugh], [sigh], [whisper], etc. (Source: BlockBeats)
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 7
  • 2
  • Share
Comment
Add a comment
Add a comment
SushiAndSlugs
· 05-26 06:41
A deep dive into emotional inline tags—can they make AI speak with an offbeat, backhanded, and vaguely sarcastic “monotone” delivery?
View OriginalReply0
FragilePosition
· 05-26 06:41
Word-level timestamps + speaker separation, the joy of podcast editing
View OriginalReply0
MempoolSparrow
· 05-26 06:35
WebSocket real-time stream $0.2/hour, is it cheaper or more expensive than Whisper?
View OriginalReply0
GateUser-b6d80ba0
· 05-26 06:30
Starlink customer service is now in use, no wonder the last time I called customer support, it felt like talking to an AI on the other end.
View OriginalReply0
AirdropMileCounter
· 05-26 06:28
25+ language coverage, how is the Chinese performance? Has anyone tested it?
View OriginalReply0
ReflectiveChainShadow
· 05-26 06:28
The same audio stack links the car system, satellites, and chat—there’s definitely something going on with this xAI ecosystem closed loop.
View OriginalReply0
MintAfterCoffee
· 05-26 06:28
What is reverse text normalization, a cutting-edge technology? Can someone knowledgeable explain it in detail?
View OriginalReply0