Blog | Soniox - NFHN Reader

Blog

What’s new

Read all the latest news and release updates from Soniox.

Soniox v5 Real-Time: Turning live conversations into structured intelligence

Soniox v5 Real-Time is now available with major improvements in accuracy, speaker separation, language ID, endpointing, translation, and more.

Soniox v5 Async: Turning real-world speech into structured data

Soniox v5 Async is now available with major improvements in accuracy, speaker separation, language ID, and context.

Partnership between Tencent Cloud and Soniox

Soniox is now natively integrated into Tencent Cloud, bringing multilingual voice AI to a wider developer audience.

Run global voice agents with LiveKit and Soniox

Soniox Speech-to-Text and Text-to-Speech APIs are now fully integrated with LiveKit. Build multilingual voice agents with a unified speech layer.

Build a real-time speech-to-speech translator with Soniox

Build a voice-in, voice-out translator using Soniox real-time speech-to-text, translation, and text-to-speech APIs.

How to build a multilingual voice bot with Pipecat and Soniox

Soniox Text-to-Speech is now fully available in Pipecat. Plug it in and build your multilingual voice bot with Soniox API.

Introducing Soniox Text-to-Speech

Generate high-fidelity speech in 60+ languages, built for names, alphanumerics, language switching, and ultra-low-latency streaming.

New Soniox SDKs

Introducing the new Soniox SDKs for Python, Node, Web, React, and React Native.

Soniox named best-in-class STT for voice agents

In a comprehensive study by Daily (Pipecat), Soniox was recognized as a top-tier provider for real-time voice agents.

app

The end of English-first voice AI, now on desktop

Work, think, and use your voice in your own language.

Soniox v4 Real-Time: The new standard for real-time speech AI

A speech recognition model purpose-built for low-latency voice interactions.

Soniox v4 Async: Human-parity speech recognition across 60+ languages

We’re launching Soniox v4 Async, a major leap forward in speech recognition.

app

Soniox App: Everything you do with voice in your language

The Soniox App is now one unified voice workspace.

app

Introducing the Soniox App for business

A shared space for your team’s calls, meetings, and interviews.

Soniox Sovereign Cloud for effortlessly global, compliant, and performant speech AI

Sovereign Cloud is now available, letting you run Soniox in regions where your users and data live.

app

Introducing the new Soniox App: transcribe, translate, and understand any speech, in any language

Soniox App brings the world’s most advanced speech AI into your pocket.

Free credits update for Soniox API

We’re discontinuing free API credits due to widespread abuse and spam signups.

Soniox v3

Soniox v3 is our most advanced speech AI yet — designed for accuracy, fluency, and speed at scale.

The fastest way to find out which voice AI actually works

Choosing a speech API isn’t just about specs. It’s about what happens when real-life audio hits the system.

Just launched: Real-time speech translation AI for the world

Today, we’re launching real-time multilingual speech translation in the Soniox API, as well as a brand new mobile app that brings the same technology to your pocket.

app

Introducing Soniox App

The first mobile app powered by a universal speech AI model that delivers real-time, high-accuracy transcription in over 60 languages.

One speech AI for the world

Soniox Speech-to-Text AI — a single, universal model that understands speech in over 60 languages, in real-time, with high accuracy. And more importantly, it works the way people actually speak.

Soniox 7B

Most powerful 7B large language model. Supports English and code with 8K context. Matches GPT-4 performance on some benchmarks. Built on top of Mistral 7B, enhanced with additional pre-training and fine-tuning for strong problem-solving capabilities.

German speech recognition AI

A small gift from Soniox before the holidays: a groundbreaking AI capable of recognizing German and English speech simultaneously with extreme precision.

Portuguese speech recognition AI

We are thrilled to announce our new groundbreaking AI that can simultaneously recognize Portuguese and English speech with extreme precision.

Soniox and Scribe elevate efficiency and accuracy of medical documentation

Soniox and Scribe are pleased to announce the integration of Soniox's groundbreaking speech recognition AI into Scribe's world-class platform and integration technologies for mobile and desktop.

Italian speech recognition AI

We are unveiling a cutting-edge bilingual AI model proficient in discerning both Italian and English with unparalleled accuracy.

French speech recognition AI

We are thrilled to introduce a pioneering AI technology capable of recognizing both French and English speech with unparalleled precision.

Spanish speech recognition AI

We are super excited to announce our groundbreaking AI that can simultaneously recognize Spanish and English speech with extreme precision.

Chinese speech recognition AI

We are thrilled to announce the release of our groundbreaking Chinese speech recognition AI, achieving unparalleled precision and comprehension of the Chinese language.

Korean speech recognition AI

We have introduced game-changing Soniox Korean speech recognition AI, capable of recognizing both Korean and English speech with extreme precision.

Soniox Speech AI achieves extreme accuracy

We have released groundbreaking speech recognition AI, achieving extreme levels of accuracy and unlocking new possibilities in human-machine interaction.

Release of Soniox 2.0: Transcribe, store, search, analyze and understand audio on one platform

Soniox Inc released an audio AI infrastructure that enables developers and businesses to transcribe, store, search and understand audio on one platform These are the core building blocks for any audio/voice/speech application.

Speech recognition benchmarks for German and Spanish March 2023

Speech recognition benchmarks comparing Soniox, Google, AWS, Azure, Rev AI, Deepgram, AssemblyAI, OpenAI, Speechmatics and NVIDIA on Spanish and German languages.

Soniox and OpenAI speech recognition benchmarks March 2023

Speech Recognition benchmarks comparing Soniox and OpenAI Whisper on 5 different real-world datasets varying in acoustic conditions, speaking styles, accents and topics in the English language.

Soniox speech-to-text benchmarks November 2022

Speech-to-Text benchmarks comparing Soniox, Google, AWS, Azure, Rev AI, Deepgram, AssemblyAI, OpenAI and Speechmatics on 4 different real-world datasets in the English language.

Soniox releases Knowledge Augmented Audio AI

Soniox has developed Soniox Knowledge Graph to further enhance the accuracy of Soniox Speech Recognition AI and to provide relevant contextual information about the recognized speech in the audio.

Soniox debuts AI Speech Recognition Platform

Soniox Inc launched the Soniox AI Speech Recognition Platform, the world’s first self-learning artificial intelligence for automatic speech recognition.