xAI launches Grok speech-to-text and text-to-speech API

robot
Abstract generation in progress
ME News Report, April 18 (UTC+8), xAI recently announced the official launch of two standalone audio APIs: Grok Speech-to-Text (STT) and Grok Text-to-Speech (TTS). Grok STT offers high-accuracy, low-latency transcription services, supporting batch processing via REST API and real-time streaming via WebSocket API, with features including word-level timestamps, speaker separation, multi-channel support, and intelligent reverse text normalization. The article mentions that in benchmark tests across multiple fields such as phone calls, meetings, videos/podcasts, its word error rate outperforms mainstream commercial models like ElevenLabs, Deepgram, and AssemblyAI. The service supports over 25 languages, priced at $0.10 per hour for batch processing and $0.20 per hour for streaming. Grok TTS can generate fast, natural, and expressive speech, supporting fine-grained control through simple voice tags, priced at $4.20 per 1 million characters. Both APIs are built on the same technology stack that powers Grok Voice, Tesla vehicle support, and Starlink customer support. (Source: InFoQ)
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
  • 5
  • 1
  • Share
Comment
Add a comment
Add a comment
GateUser-99725296
· 12h ago
Can this price volatility be sustained? Let's see how Azure and AWS respond.
View OriginalReply0
0xLateBreakfast
· 12h ago
Batch and streaming are double the price; the business needs to do a good job of calculating the costs.
View OriginalReply0
RiskOffRina
· 13h ago
Multi-channel support is crucial for conference transcription; finally, no need to align manually.
View OriginalReply0
WalletHealthInspector
· 13h ago
Grok Voice uses these two APIs, right? Finally separated out.
View OriginalReply0
MosaicBowtieRealm
· 13h ago
What specific parameters can fine-grained control tags manage? Is it speech speed, pitch, or emotion?
View OriginalReply0