Gladia Introduces Solaria: An AI-Powered Multilingual Speech Recognition Model for Speech-to-Text Transcription

Gladia, a notable player in the AI transcription and audio intelligence sector, recently introduced Solaria, a cutting-edge automatic speech recognition (ASR) model. This innovative tool is designed to transform real-time communications for call centers and various voice-focused platforms.

With Solaria, businesses can improve their customer service capabilities through advanced AI voice technology. This technology boasts support for over 40 languages, a feature that was previously unavailable with many existing solutions, all without sacrificing quality or speed.

In the call center industry, outsourcing has long been a method for cost reduction. However, a new challenge has arisen: the need for seamless and scalable multilingual support. A staggering 49% of global executives report experiencing financial losses due to language barriers, underscoring the urgent demand for high-quality, scalable multilingual solutions.

Jean-Louis Queguiner, CEO of Gladia, emphasized the increasing importance of voice AI, stating, “We’ve witnessed a remarkable rise in voice technology, and with the introduction of Solaria, we are bringing a real-time model with advanced capabilities to market.” He emphasized that Solaria aims to be the fastest and most accurate solution available, promising coverage for 100 languages.

Solaria: A Global Customer Experience Solution

Solaria is specifically crafted as a speech-to-text (STT) engine optimized for global scalability. It addresses the needs of contemporary contact centers, where both AI automation and human agents rely on accurate, low-latency, and real-time support to be successful.

The model has achieved impressive benchmarks, achieving a Word Accuracy Rate (WAR) of 94% in frequently spoken languages such as English, Spanish, and French. Its processing speed is notably fast, with an ultra-low latency of just 270 milliseconds, allowing conversations to flow naturally.

Speed is often prioritized in real-time speech-to-text solutions; however, Solaria recognizes the equal importance of accuracy and language diversity. Unlike many other models that focus solely on speed, Solaria balances speed and precision while also offering unmatched language support—covering 100 languages and providing exclusive capabilities for 42 languages that competitors do not support. This is especially beneficial for heavily populated regions and key outsourcing countries like Bangladesh, India, and the Philippines, where native-level accuracy in regional languages is crucial.

With its features including native transcription, real-time code-switching, and translation in all supported languages, Solaria enables businesses to explore global markets without limitations.

Some essential features of Solaria include:

• Top-tier accuracy in high-population languages such as Tagalog, Bengali, and Urdu.
• Model adaptability for industry-specific terminology to extract important data like names or addresses.
• Adaptive speech processing that ensures high accuracy in the often noisy environment of call centers.
• Enterprise-grade data security, compliant with GDPR, HIPAA, and SOC 2 standards.

By incorporating Solaria into their offerings, Gladia enables businesses to boost customer service efficiency with improved AI-powered voice agents. This also enhances the reliability of Interactive Voice Response (IVR) systems and virtual assistants across multiple languages and streamlines human-assisted workflows via real-time transcriptions and translations.

As Jean-Louis Quéguiner remarked, “Speech is the natural way for humans to connect with the world. Automated speech recognition is bridging the gap, allowing humans and AI to effectively communicate.” With Solaria, Gladia aims to revolutionize AI voice technology, driving efficiency while creating more impactful customer experiences in diverse languages and markets.

Gladia currently serves over 700 enterprise clients worldwide, including industry leaders like Attention, Circleback, and VEED.IO. They offer robust service and scalability, supported by dedicated infrastructure in the U.S. and Europe, ensuring reliable performance for critical applications. Businesses looking to expand globally and enhance their customer experience can use Gladia’s API to get started.

As part of the launch of Solaria, Gladia has teamed up with LiveKit, a prominent framework for developing real-time AI voice agents. This partnership provides developers with global language capabilities through seamless integration with Gladia’s API.

After securing $16 million in Series A funding in 2024 and launching Solaria, Gladia is solidifying its position as a leading provider of comprehensive API audio infrastructure, combining speech recognition, generative AI, and voice generation to help enterprise users maximize real-time audio data.

Founded in 2022 in Paris by Jean-Louis Queguiner and Jonathan Soto, Gladia has already made significant inroads with over 150,000 users and a client roster that includes major entities such as Circleback and Method Financial.

Voice activation produces a response with a 300-millisecond delay, and transcription takes an additional 100 milliseconds, resulting in near-instant results for users.

To enhance accuracy further, Queguiner notes that the company is focusing on training with more data and making the current data more robust. Although enterprise pricing details have not been released, he assures it will remain competitive. The company currently employs nearly 40 staff members.