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 independent audio APIs: Grok Speech to Text and Grok Text to Speech. Both come from the same audio stack that supports Grok Voice, Tesla’s in-car system, and Starlink customer service. This time, they are opened up in the form of 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, with millisecond-level return; the WebSocket API is designed for real-time audio streams. Included capabilities cover word-level timestamps, speaker separation (diarization), multi-channel recognition performed separately for each channel, and Inverse Text Normalization, which automatically reshapes spoken numbers, dates, and currencies from conversational speech into standardized structured text. The language coverage spans 25+ languages and allows seamless switching within conversations.

xAI also released a set of word error rate (WER) comparisons (the lower the better): overall scenario—Grok 6.9%, ElevenLabs 9.0%, Deepgram 11.0%, AssemblyAI 12.9%. The gap is even larger for “telephone call entity recognition”: Grok 5.0%, with the corresponding three at 12.0%, 13.5%, and 21.3%, respectively. In common business scenarios such as meetings, video podcasts, and phone calls, Grok is also slightly ahead across the board. These figures are published by xAI based on its own testing, and there has not yet been any third-party re-verification.

On pricing: STT batch processing is $0.10 per hour, streaming is $0.20 per hour; TTS is $4.20 per 1 million characters. TTS supports inline Speech Tags to control emotion and intonation, such as \[laugh], \[sigh], \[whisper], and \[.] (Source: BlockBeats)