Meraki — Web and Desktop Platform for a Children's Language School
We built and keep growing Meraki, the platform behind a language school for children: a web app for lessons and voice exercises, and a front-desk desktop app wired to a time clock, a receipt printer and a TV display, all synced with the cloud. It serves 2,600+ students and has logged 150,000+ sessions since September 2025.
The Problem
Create a fun and effective language-learning experience for children, combining speech recognition, gamification, and real-time feedback — all while maintaining cross-browser stability and a secure CI/CD pipeline.
Our Solution
We developed an AI-driven learning ecosystem using a Laravel backend, React front-end, and a Python microservice for pronunciation scoring. The system supports audio exercises, real-time evaluation, and dashboards for students, parents, and teachers. Infrastructure was containerized, secured via Nginx and Cloudflare, and deployed with full observability on DigitalOcean.
Architecture
The Impact
100+ students practising at the same time at peak, each one sending recordings to the pronunciation-scoring API
2,600+ students, 780+ parents and 19 teachers on the platform, with 870 active subscriptions
150,000+ sessions and 46,000+ hours of learning since September 2025, about 18 minutes per session
1.4M+ page views and 10,000 users from across eastern Algeria (Google Analytics)
Up to 10,000 voice recordings submitted in a single day
Subscriptions, payments, dues and discounts run from the platform instead of spreadsheets
Automated builds and zero-downtime deployments while the school keeps running
Key Features
Students: structured lessons, voice repetition exercises, XP, seasons and leaderboards
Front desk: a desktop app connected to a time-clock terminal, a receipt printer and a TV that shows each arriving student's subscription status and session room
Cloud sync: attendance, payments and receipts from the front desk sync with the web platform
Parents: progress reports, attendance and absence alerts, subscription and payment status
Teachers: groups, timetables, live sessions, attendance and audio review
Finance: subscription plans, payments with receipts, outstanding dues, discount rules, expenses and salaries
Admins: courses, course runs, grade levels, groups and reporting dashboards
Voice AI: pronunciation scoring, accuracy feedback and recording analytics
AI assistant access: an MCP server lets the school's administrators query and manage the school from Claude
Challenges
Ensuring consistent audio capture across mobile browsers
Handling noise and accent variation in pronunciation scoring
Delivering safe, zero-downtime deployments
Managing CORS and secure media permissions in multi-device environments
Lessons Learned
Establish measurable latency and accuracy goals for AI features
Validate voice capture quality early in the design phase
Prioritize observability and reproducibility in deployment scripts
Engage real users early to refine evaluation feedback
Our Process
Discover
2 weeksScope
4 weeksBuild
14 weeksLaunch
2 weeksEvolve
ongoingProject Gallery
Want to be the next success story?
Let's discuss your project and turn it into an amazing digital reality