Self-conducted reading assessments for any language
Readup supports large-scale reading assessments at a fraction of the cost and time of traditional methods.
Built with researchers at the University of Cape Town and Western Sydney University, and deployed on Gates Foundation–funded projects.
How it works
The Readup Platform
Readup is an AI-powered platform that provides child-directed reading assessments. Using advanced speech recognition, it automates reading diagnostics, making reading assessments accessible and affordable for children worldwide.
Our platform streamlines the process of creating, conducting, and analyzing reading assessments, making it possible to reach learners in any language, anywhere in the world.
- Create Assessments. Easily design reading assessments for various skills, including phonics, fluency, and comprehension. Our platform supports multiple languages including English, Xhosa and Sepedi.
- Streamlined Data Collection. Collect audio data using any Android device. Our self-guided interface allows learners as young as five years old to complete assessments independently, eliminating the need for trained personnel and making large-scale data collection much faster and cheaper.
- Flexible Assessment Analysis. Get an accurate assessment of students' reading skills by analyzing the collected data with either human evaluation or our advanced AI system.
Features
Everything you need to run reading assessments
Self-guided app
Our app works on low-end devices and is so simple that 5-9 year olds can use it unaided to complete diagnostic, formative or summative assessments.
Offline Data Collection
Reading assessment data and audio responses are collected offline without special equipment and synced to the cloud when internet access is available.
Multi-language support
Assessments can be created for any language, with AI automated marking models that support African languages like Xhosa and Sepedi.
Human & AI Assessment
Our web interface allows reading assessments to be marked by trained markers or AI models from anywhere. Our AI models are trained by experienced EGRA markers and are calibrated against these markers.
Case Study
EGRA-AI powered by Readup
Reliable reading assessments like EGRA (Early Grade Reading Assessment) give a trustworthy picture of a child's reading skills — but run one-on-one, they're complex, expensive, and time-consuming. Readup set out to change that: automating EGRA in African languages so assessment can move from one-to-one to one-to-many, at a fraction of the cost.
- 4,500+
- Learners assessed
- 240,000+
- Audio recordings collected
- 500,000+
- Human labels created
- 2
- African languages
2023 · Field trial
Proving it in the field
In July 2023 the Gates Foundation funded a project led by Dr Cally Ardington of the University of Cape Town to automate EGRA in African languages — EGRA-AI.
The Readup-powered app was deployed by EGRA agents into 120 schools taking part in a field trial with Funda Wande, a not-for-profit equipping teachers to teach reading-for-meaning and calculating-with-confidence in Grades R–3. Over 2,000 learners completed assessments, laying the foundation for the next generation of EGRA-AI.
2024 · Learning by Doing
Expanding to isiXhosa
With a Learning by Doing research grant from AI-for-Education, the EGRA-AI Consortium expanded the work to isiXhosa. We rebuilt the app around a tap-to-speak interface with built-in noise detection, worked with local language experts to design the assessment and quality-assure marking, and built a platform for mother-tongue markers to label recordings captured in the field.
Across three waves, 148,962 recordings were collected from 2,796 Grade 1–4 learners. Every recording was labelled by three independent mother-tongue markers — over 500,000 labels in 2024 — and only consensus labels were used to train custom marking models built on wav2vec 2.0. The findings were published in peer-reviewed research at AIED 2025.
The results
AI marking that matches human markers
One hundred assessments were held back from training to test the models — and self-administered EGRA-AI scores were validated against traditional one-on-one EGRA (a correlation of 0.84 across 345 learners who completed both).
95%
AI marking accuracy
Item-level accuracy on recordings where all three human markers agreed (91% across all items).
0.99
Correlation with human markers
AI assessment scores explain 98–99% of the variation in human marker scores — with no accuracy difference between boys and girls.
<1c
AI marking cost per learner
Versus $6.94 for triple human marking — cutting the all-in cost of a full assessment roughly in half.
8×
Learners per field worker
One facilitator runs eight self-guided assessments at once — children simply sit with a tablet and headphones.
Published research
AIED 2025An End-to-End Approach for Child Reading Assessment in the Xhosa Language
Chevtchenko, Navas, Vale, Ubaudi, Lucwaba, Ardington, Afshar, Antoniou & Afshar
Our team's peer-reviewed paper presents a novel dataset of Xhosa child speech from the EGRA-AI project — labelled by multiple mother-tongue markers and validated by an independent EGRA reviewer — and benchmarks three fine-tuned state-of-the-art speech models: wav2vec 2.0, HuBERT, and Whisper.
Our Team
Meet the team building Readup
Expert in computational architectures and algorithms for vision, memory, and auditory sensing systems.
Project manager and startup founder with a background in tech and enterprise solutions.
Dr Soheil Afshar
Clinical Neuropsychologist
Developmental Paediatrics
Expert in learning difficulties supporting children with reading/literacy, numeracy, and attention problems.
Specializes in neuromorphic systems and computational neuroscience, focusing on event-based vision and processing.
Dr Claire McAulay
Senior Clinical Psychologist
Black Dog Institute
Expert psychologist working with adolescents and adults, specializing in evidence-based therapies like cognitive behavioural therapy.
Expert software architect specializing in designing scalable and efficient systems.
Expert software engineer with a passion for developing innovative technological solutions.
Speech scientist researching speech perception and language learning across languages and age groups.
Work with us
Let's collaborate
We partner with NGOs, universities, and governments to run large-scale reading assessments for children. Our approach involves working closely with your team to customize assessments that are aligned with the specific languages and contexts of the children you serve.
Throughout the collaboration, we focus on refining the process together, offering a seamless integration with your existing systems. Whether for diagnostic, formative, or summative assessments, we combine automated and human marking to provide efficient and accurate results.