Voice AI
for Literacy

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.

Our speech recognition technology delivers accurately and safely for each stage of the journey and for all use cases, from literacy practice and assessment to screening.

early literacy letters

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.

early literacy phonics

Phonemic awareness

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.

early literacy phonics words


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.

early literacy sentences

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.

Use cases

Phonemic awareness

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.

Early reading

Reading tools powered by SoapBox return feedback and insights to guide teachers’ instruction.

We deliver word accuracy assessments down to the phonemic level in real-time to the teacher dashboard.

Dyslexia screening

Our voice engine powers dyslexia and other screeners, incorporating oral responses into screening assessments to identify a child’s oral language skills and challenges.


With SoapBox, voice-enabled literacy activities such as multiple choice questions can simultaneously analyze 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 and omissions and accuracy measures like Words Correct Per Minute (WCPM) and prosody are returned as real-time feedback to teachers.

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, CTO, SoapBox Labs