LotsDirectoryLotsDirectory

Soniox

One speech AI API for realtime speech-to-text, text-to-speech, and translation in 60+ languages.

Visit website

Description

One speech AI API for realtime speech-to-text, text-to-speech, and translation in 60+ languages. Low-latency streaming, native-speaker accuracy, and seamless language switching for live voice applications.

About Soniox

What is Soniox?

Soniox is a real-time voice AI platform providing a unified API for speech-to-text, text-to-speech, and speech translation across 60 languages.

What Soniox does

The platform streams real-time speech transcription, translates across 3600 language pairs without waiting for sentence boundaries, generates synthesized speech from text, and clones voices from audio reference clips.

Who Soniox is for

Developers and enterprises building voice agents, meeting transcription systems, dictation tools, and real-time translation applications.

How Soniox works

Developers integrate the Soniox API to stream audio for transcription and translation in a single call, or submit text to generate streaming audio output with character timestamps.

Key features

  • Transcribes real-time and recorded speech across 60 languages with speaker detection.
  • Translates speech across 3600 language pairs during live conversations.
  • Generates expressive text-to-speech audio in 60 languages with instant voice cloning.
  • Handles mixed-language conversations where speakers switch languages mid-sentence.
  • Captures alphanumeric characters including phone numbers, IDs, and email addresses.
  • Provides global data residency with deployments in the US, EU, Japan, and India.

Pricing

As listed on soniox.com, checked 8 November 2026. Prices can change; see the site for current terms.

  • Async speech-to-text: $0.10 per hour for recorded audio file transcription.
  • Real-time speech-to-text: $0.12 per hour for streaming transcription.
  • Text-to-speech: approximately $0.70 per hour of generated speech audio.
  • Speech translation: included in real-time speech-to-text calls at no extra charge.

Facts from soniox.com, checked 8 November 2026

  • Combines speech-to-text, text-to-speech, and translation across 60 languages in one API.
  • Translates across 3600 language pairs in real time without waiting for sentence ends.
  • Processes audio in memory without storing raw customer voice data.
  • Offers regional data residency deployments in the US, EU, Japan, and India.

Written from Soniox's own website (Homepage, Pricing, Features, About). The maker can claim this listing to edit it.

Categories

Developer ToolsAI & Machine LearningMobile Apps

Frequently asked questions

What speech capabilities does the Soniox API combine?

The API combines real-time speech-to-text, text-to-speech, and live speech translation in a single interface.

How does Soniox handle conversations with multiple or mixed languages?

The unified model detects mid-sentence language switches automatically across 60 languages without manual configuration.

Does speech translation require a separate API call?

No. Speech translation is built directly into speech-to-text and returns translated text in the same real-time call.

What security and compliance standards does Soniox maintain?

Soniox complies with SOC 2 Type 2, ISO/IEC 27001:2022, HIPAA, and GDPR standards with in-memory audio processing.

Listed from the product's public website. Own this product? Claim this listingRequest removal