xAI Grok Speech to Text and Text to Speech APIs

xAI has launched two standalone audio APIs: Grok Speech to Text (STT) and Grok Text to Speech (TTS). These endpoints allow developers to integrate high-quality audio transcription and voice generation into applications such as voice agents, real-time transcription tools, and accessibility solutions.

Grok Speech to Text (STT)

High-Accuracy Transcription and Low Latency

Grok STT provides transcription capabilities via two primary interfaces: a REST API for processing large audio files in milliseconds and a WebSocket API for real-time, low-latency streaming transcription. The system includes word-level timestamps, multichannel support, and speaker diarization to identify different speakers in both pre-recorded and streaming audio.

Inverse Text Normalization

Unlike models that provide raw spoken words, Grok STT utilizes Inverse Text Normalization to convert spoken language into structured output. This process intelligently formats numbers, dates, currencies, and other entities to ensure the output is professional and readable.

Performance Benchmarks

Grok STT is designed for enterprise-grade transcription, specifically excelling in entity recognition for medical, legal, and financial business use cases. In comparative Word Error Rate (WER) testing across various domains, Grok STT demonstrated the following performance:

Domain (Word Error Rate) Grok STT ElevenLabs Deepgram AssemblyAI
Phone Call Entities 5.0% 12.0% 13.5% 21.3%
Video/Podcasts 2.4% 2.4% 3.0% 3.2%
Meetings 10.9% 12.2% 16.3% 15.7%
Telephone 9.3% 9.4% 11.0% 11.2%
Overall 6.9% 9.0% 11.0% 12.9%

Multilingual Support and Pricing

The API supports over 25 languages, allowing for seamless language switching. Pricing for Grok STT is set at $0.10 per hour for batch processing and $0.20 per hour for streaming.

Grok Text to Speech (TTS)

Natural Voice Generation

Grok TTS enables the conversion of long-form text into speech using a REST API, or real-time speech generation via a WebSocket API. The model is designed for natural and expressive delivery.

Fine-Grained Prosody Control

Developers can control the emotion and prosody of the generated speech using inline and wrapping speech tags. Supported tags include [laugh], [sigh], [whisper], <emphasis>, <slow>, and <pause>, allowing for the creation of lifelike audio without complex markup.

Pricing

Text to Speech is priced at $15.00 per 1 million characters.

Sources

Related

  • Dispatch
    xAI Grok Voice Agent API Release

    xAI has launched the Grok Voice Agent API, a high-speed, multilingual voice agent system that ranks #1 on Big Bench Audio and costs $0.05 per minute.

  • Dispatch
    Grok Voice Think Fast 2.0 Release Notes

    xAI has released Grok Voice Think Fast 2.0, a next-generation voice model featuring improved intelligence, transcription accuracy, and conversational efficiency.

  • Dispatch
    xAI Grok Voice Think Fast 1.0 Release

    xAI has released grok-voice-think-fast-1.0, a flagship voice model designed for complex enterprise workflows, high-volume tool calling, and low-latency real-time reasoning.

  • Dispatch
    xAI Custom Voices Release

    xAI has introduced Custom Voices, allowing users to clone their own voice from a short recording for use across Grok Text to Speech and Voice Agent APIs.

  • Dispatch
    xAI Grok 4 Release Notes

    xAI has released Grok 4, a model featuring scaled reinforcement learning, native tool use, and a high-performance 'Heavy' variant that achieves state-of-the-art results on ARC-AGI V2 and Humanity's Last Exam.