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

ME News, April 18 (UTC+8): According to Beating Monitoring, xAI has launched two separate audio APIs—Grok Speech to Text and Grok Text to Speech. Both are built on the same audio stack that supports Grok Voice, Tesla’s in-car system, and Starlink customer service. This time, the capabilities are released as standalone endpoints, enabling developers to directly integrate applications such as voice agents, real-time transcription, accessibility tools, and podcasts.

STT offers two modes. The REST API is for batch transcription of large audio files, returning results at millisecond-level speed; the WebSocket API is designed for real-time speech streams. Included capabilities include word-level timestamps, speaker diarization, multi-channel recognition per channel, and Inverse Text Normalization—automatically converting spoken numbers, dates, and currencies into standardized structured text. Language coverage exceeds 25 languages, with seamless switching during conversations.

xAI also released a set of word error rate (WER) comparisons (the lower the better): for the overall scenario, Grok is 6.9%, ElevenLabs 9.0%, Deepgram 11.0%, and AssemblyAI 12.9%. The gap is even wider for “telephone call entity recognition”: Grok is 5.0%, while the corresponding figures for the other three are 12.0%, 13.5%, and 21.3% respectively. In typical business scenarios such as meetings, video podcasts, and phone calls, Grok also maintains a slight lead. These figures are self-tested and published by xAI, and there has been no third-party re-testing yet.

In terms of pricing, STT batch is $0.10 per hour, and streaming is $0.20 per hour; TTS costs $4.20 per 1 million characters. TTS supports inline Speech Tags to control emotion and intonation, such as \[laugh\], \[sigh\], and \[whisper\], and \[...\].

（Source: BlockBeats）