A voice-powered journey from phonics to fluency
Accurate, voice-enabled solutions give teachers more time to personalize their instruction and support students of every age and literacy level – whether they’re emergent readers or building more fluent reading skills – and in every educational setting, from busy classrooms to remote learning at home.
Our speech recognition technology delivers accurately and reliably for each stage of the journey and for all use cases, from literacy practice and assessment to screening.
Letter names and sounds
Our voice engine supports a child’s learning and assessment from the earliest activities, such as letter naming or letter sounds. We can return highly accurate scoring data for these short sounds.
A foundational step in literacy development. Whether blending, decoding, segmenting, or isolating, our engine supports learning and assessment items for phonemic awareness and phonic scoring.
Whether assessing real or nonsense words, our voice engine returns scoring at both word and phoneme level, offering teachers highly accurate and granular data to measure a child’s progress.
Sentences and passages
As a child reads leveled decodable readers and informal texts, our voice engine returns scoring at sentence, word, and phoneme level to accurately assess their progress.
Our voice engine powers phonological and phonemic awareness skills by listening to, capturing, and analyzing a child’s short utterances, such as letter names and sounds. Download our product sheet to learn about these and the other phonemic awareness use cases we support.
Reading tools powered by SoapBox Educate return scores and rewards to students while demonstrating progress to teachers.
We deliver word accuracy assessments down to the phonemic level, and right into your teacher’s dashboard.
With SoapBox Educate you can voice-enable literacy activities such as multiple choice questions and simultaneously score each response from a child against an unlimited number of words and phrases.
Reading fluency assessment
As the child reads out loud, their audio is analyzed by our voice engine.
Key metrics like insertions, hesitations, omissions and Words Correct Per Minute (WCPM) are calculated and scored for accuracy against a given text.
To pronounce the word ‘green,’ a kid might go through an entire evolution from ‘bee’ to ‘been’ to ‘breen’ to ‘green.’ We see these voice patterns in the data we use to build our models too.Dr. Amelia Kelly, VP of Speech Technology, SoapBox Labs