LanguaMetrics, Inc. is a world leader in the AI powered analysis of both Children’s and Adult’s speech.

Join us on our voice-powered journey from phonics to fluency to understandability! We have helped millions of learners—both young and old, from Beijing to Bloomington—improve their English reading, their speaking, and help qualify their English in the workforce.

Follow our decade-long timeline of producing cutting-edge speech powered products for major companies and institutions around the world. 

2012
LanguaMetrics Inc. becomes a licensee of SRI International, the world leader in speech recognition research and the original creators of Siri, later to be purchased by Apple, Inc. LanguaMetrics leverages this technology into the early literacy and English language learner spaces.
SRI international
2012
LanguaMetrics Inc. invited to present at a global conference of education publishers interested in our technology's ability to score children’s speech for pronunciation accuracy.
global conference
2012
LanguaMetrics Inc. is invited to Beijing, China to begin research and development on products for both children and adult English language learners.
global conference
2013
LanguaMetrics Inc. is invited by the Chinese Teacher’s Union to create the first National Spoken English Competition, computer-based-examination. This was the first AI-driven test to measure the speaking ability of Chinese English teachers at the University level.
chinese teacher unions
2014
LanguaMetrics Inc. partnered with Metrics China to produce a leading edge learn-to-read software for young Chinese children attending after-school programs.
metrics china
2014
LanguaMetrics Inc. begins research and development with NWEA of Portland, Oregon, creators of MAP Growth, to determine the feasibility of measuring oral reading fluency via computer-based assessment.
nwea
2017
NWEA launches MAP Reading Fluency, powered by LanguaMetrics' speech analytics and AI technology. This is the first product to market in the arena of measuring oral reading fluency via computer, thereby relieving tens of thousands of teachers from sitting knee-to-knee with students, measuring their oral reading fluency with a stopwatch.
map reading fluency
2018
NWEA’s MAP Reading Fluency, powered by LanguaMetrics, wins the prestigious CODiE Award.
nwea map reading fluency
2018
LanguaMetrics goes beyond SRI speech technology, developing its own proprietary AI and speech analytics to provide increased accuracy and develop advanced speech services for new markets.
LanguaMetrics
2018
LanguaMetrics is invited to compete to provide the call center industry’s next generation Understandability measurement for offshore customer service agents.
call-center
2019
The LanguaMetrics Understandability Measurement System is selected for use in the screening of potential offshore customer service agents. This system is the first completely AI-powered measure of Understandability. It is now in use in many of the world’s largest call center operations.
AI powered
2020
LanguaMetrics builds on its experience producing products around the world and releases FluencyRev™ for the U.S. K-5 market. The first AI-powered product with hundreds of speech-enhanced books designed as a tool for teachers in accordance with the Science of Reading.
FluencyRev
2021
LanguaMetrics begins research with the British Council to explore next-generation applications of AI in identifying areas of improvement for English language learners' speaking ability.
British council
2022
LanguaMetrics builds on the success of FluencyRev™ and releases PhonicsRev™ and EnglishRev™ into its Early Adopter program, thereby creating the RevLearning Suite!
RevLearning Suite

What is the LanguaMetrics difference?

Bespoke Artificial Intelligence

We customize our AI models to meet the needs of each application, in each market. This takes time and effort. We do not license other organizations’ technology; we roll our own. Our data scientists are constantly adapting our base models into product-specific models for a wide variety of markets.

Human Rating

The purpose of AI models is to transform human intelligence into machine intelligence. To do this with precision it is necessary to have a machine learning operations workflow, MLOPS. For a decade, LanguaMetrics has maintained a constant pipeline of classified child and adult audio recordings and hand-scored by our trained raters, as well as our partners' raters.

Machine Learning

Our data science team takes in the audio files that have been rated by our human raters to begin the model building and validation process. Our data scientists strive to constantly improve the overall accuracy of our models and quantify the results in industry standard terminology such as Precision, Recall, F1, false positives and false negatives. In AI, accuracy must be quantified across several measures.

Digital Signal Processing

Our audio and electronic engineers are pioneers in deciphering the speech signals. We build proprietary AI models to measure signal features related to background noise so that we can identify recordings that should not be scored because the results could be unfair. We use our signal expertise to expose what are known as “features” so that our data scientists are able to detect patterns that otherwise would not be visible in the data.

Phoneme Level Scoring

All results at the application level depend on our phoneme level scoring capabilities. This can cause information overload if not properly managed. In all our products across the globe, we do analysis at the phoneme level, assigning a numeric score to the utterance's pronunciation accuracy. With these results we have developed an extensive inventory of proprietary algorithms to make scoring and feedback decisions at the word and sentence level based on the phoneme level data

Accents, Equity, and Diversity

Off-the-shelf AI models and speech recognition libraries may be biased because they are based on the most readily available speech data from the internet, which is not sufficiently ethnically nor geographically diverse. Our speech data is collected from the groups of users who use our products, and we carefully monitor its diversity of accent, ethnicity, and gender. Our AI models are held to the highest standards in education and the workforce to ensure scoring results are unbiased and fair to an increasingly diverse global community.

Subject Matter Expertise

Our subject matter experts ensure that the output from our artificial intelligence and speech scoring algorithms is fit for purpose. This ensures our products are not solutions in search of problems to solve. We start with the specific challenge and then custom craft each solution. Our solutions range from the Education space to the Call Center industry and beyond.

Come join us!

Rev-up with the RevLearning Suite.

Come join us! Rev-up with the RevLearning Suite.