Enterprise-grade
speech to text transcription

High-accuracy speech-to-text for phone calls, meetings, videos, and podcasts.

Try it live

Speak into your mic and watch the transcript appear in real time.

  • Batch for files, WebSocket for low-latency streaming.
  • Word-level timestamps, diarization, and multichannel support.

Click “Start Transcribing” to begin...

Fast, accurate transcription

Grok Speech to Text is optimized for tough conditions: telephony, noise, and more.

  • Leads overall accuracy across enterprise domains.
  • Inverse text normalization for numbers, dates, and currencies.

Accuracy (higher is better)

Grok Speech to TextSpaceXAI
93.1%
Scribe v2ElevenLabs
91.0%
Nova 3Deepgram
89.0%
Universal 3 ProAssemblyAI
87.1%

Simple, transparent pricing

Usage-based billing with no hidden fees.

  • A fraction of competitors' prices.
  • Pay only for what you use.

Cost per hour of audio

Grok Speech to TextSpaceXAI
$0.10
Universal 3 ProAssemblyAI
$0.21
Scribe v2ElevenLabs
$0.22
Nova 3Deepgram
$0.26

Batch + streaming

REST API for transcribing file uploads, WebSocket for live transcription.

Drop audio to transcribe
meeting.mp3

How can I help you today

Inverse text normalization

Spoken numbers, dates, and currencies become written form.

Spoken
one hundred dollars due june twelfth at three thirty p.m.
Transcribed
$100 due June 12 at 3:30 PM

Keyterm boosting

Bias towards your product names and proper nouns.

"keyterm": [
"Grok",
"SpaceXAI",
"Understand The Unive
]

Word-level timestamps

Per-word start and end times for karaoke-style captions.

Nice0.92s
to1.14s
meet1.22s
you,1.44s
Jane.1.60s

Speaker diarization

Word-level speaker IDs to separate transcription for up to 20 speakers.

Speakers
A
Thanks for calling. How can I help?
I can't log into my account.
B
A
What's the email on file?

Multichannel

Transcribe up to 2 channels separately at the same time.

Agent
Customer

Multilingual

Enterprise-grade transcription for 25+ languages.

Language25+
EnglishSpanishFrenchGermanJapanesePortugueseItalianKoreanHindiArabicDutchTurkish

Filler word control

Strip uh, um, and er from text and the words array by default.

Spoken
I, uh, need the, um, invoice
Transcribed
I need the invoice

Explore the full Grok Voice suite

Speech to Speech

Intelligent, real-time conversations with sub-second latency.

$0.05 / min

Speech to Text

Enterprise-grade transcription for meetings, videos, and more.

$0.10 / hr batch$0.20 / hr streaming

Text to Speech

Natural, expressive speech from text in 25+ languages.

$15.00 / 1M characters

Voice Cloning

Clone a voice from two minutes of speech and use it instantly.

30 Free Clones

Live Translation

Speak in one language and be heard in another.

Coming Soon