Nitish Prasad

Native iOS / offline first / product restraint

Sprout Math

Children need clear math practice, and parents need control over their data. I built Sprout Math around short offline sessions, on-device progress, and predictable hints.

Native iOS productOffline practice
Customer
Children and parents
Scope
Independent 0→1 build
Platform
Native iOS
Privacy
No accounts
Open live site ↗ View GitHub ↗ Visual worlds Offline system Privacy and terms
Decision
Keep progress, content, hints, and diagnostics on device with no account requirement.
Why
A child-focused first release needs predictable help, fast sessions, and a small privacy surface.
Result
An offline native math product with a working practice loop and local progress.
Quick readRestraint

Product restraint carries the value.

Kids apps lose trust when they lag, require accounts, push subscriptions, or bury parents under settings. Sprout Math keeps the core loop local, the hints deterministic, and the session design short enough to remain calm.

Focused learning loop

Short sessions, adaptive placement, visible progress, and deterministic correction after repeated misses.

Parent trust first

Parent-gated settings, local diagnostics export, no ads, no accounts, no surprise network dependency.

Content model

A bundled K–5 practice curriculum, deterministic hints, narration, and six visual worlds. See the live product site for the current content inventory.

Architecture

Core Data persistence for profiles, attempts, mastery, history, and progress state vectors.

VisualsEvidence

The look is playful. The system underneath is controlled.

Candyland visual world
Candyland
Axolotl Lagoon visual world
Axolotl Lagoon
Rainbow Unicorn visual world
Rainbow Unicorn
Stars and Space visual world
Stars and Space
Superhero City visual world
Superhero City
Turbo Cars visual world
Turbo Cars
ArchitectureOffline

Offline-first product architecture is a trust decision.

Local state

Core Data

Profiles, attempts, mastery, lesson state, stickers, and history remain on device.

Learning loop

Deterministic hints

3-tier hint logic and correction flow avoid opaque tutoring behavior.

Content

Template bank

Question variety comes from maintained templates, not remote generation.

Trust

No account loop

Children can practice without accounts, ads, analytics, or subscriptions.

DecisionsScope

Most of the hard decisions were about boundaries.

Short sessions over endless play.

Practice stays repeatable and parent-friendly.

Universal iOS, portrait lock.

Reducing layout modes protects quality in a first version.

Deterministic hints.

Kids products need predictable help before cleverness.

Local export over remote analytics.

Diagnostics can exist without turning a child product into a tracking surface.