Machine Intelligence Digital Assistant

One safe intelligence.
Every learning experience.

MIDA is the universal educational backend that helps every Sky Education app understand the learner, explain at the right level, and continue one trusted learning relationship.

Child-firstModel-independentAPI-firstEnglish + 中文
MIDA Core 0.2BILINGUAL FOUNDATION
A learner asks

“Why does the Moon seem to follow us?”

01Understand02Protect03Explain04Check

Not another chatbot.

A shared intelligence layer for an entire educational ecosystem.

Live experience

Ask once. Keep going until it makes sense.

MIDA changes the explanation—not the truth—according to the learner’s age, language, and request.

M
MIDA Safety layer active
LIVE DEMO
M

Hi! I’m MIDA, an AI learning friend from Sky Education. Ask me something, and we can make the explanation simpler or deeper until it clicks.

Transparent AI · child-safe introduction

Presentation mode uses a limited verified knowledge set. MIDA says when it cannot verify an answer.

The explanation ladder

Understanding has more than one doorway.

A child can ask MIDA to make an answer shorter, easier, more visual, more detailed, or practical. MIDA keeps the same concept while rebuilding the path toward it.

Quick answerMake it easierShow the stepsGive an exampleGo deeperLet me try
Same question3 useful depths
Simple

A shadow appears when something blocks light.

Age 6–8
Clear

Light travels in straight lines. An object blocks some rays, leaving a darker area behind it.

Age 9–12
Deep

The light source’s size and position create umbra and penumbra regions with different light intensity.

Advanced

Safety is architecture

Protection is not a warning added at the end.

It is part of how MIDA receives, reasons, responds, remembers, and escalates.

01

A friend, never a replacement

MIDA is warm and consistent, but always says it is an AI. It never asks for secrecy or tries to replace family and real relationships.

02

Privacy by design

Educational progress can be remembered. Sensitive details are minimized, redacted, and never collected just to make conversation feel personal.

03

Safety before generation

Every request passes an age-aware safety check before any model or knowledge provider receives it. Every answer is checked again before delivery.

04

Truth over confident guessing

MIDA separates what it knows from what it cannot verify and asks for context when accuracy depends on missing information.

Universal by design

One backend. Many educational worlds.

Applications keep their own experience. MIDA provides the intelligence, safety, continuity, and learning logic underneath.

01

Connected apps

Reading · speaking · homework · science · future products

02

MIDA gateway

One stable, versioned API for every educational experience

03

Child-safety layer

Age policy · privacy guard · risk handling · response check

04

Teaching engine

Explanation depth · examples · questions · understanding checks

05

Knowledge & models

Approved content and interchangeable intelligence providers

06

Learning memory

Progress · preferences · misconceptions · next best action

API-first foundation

Connect a new learning app without rebuilding its intelligence.

A versioned contract carries learner context, explanation preferences, safety decisions, and pedagogical metadata in one response.

REST API v1Tenant-awareProvider-neutralStructured metadataBilingual
Open developer documentation
POST /api/v1/chat
{
  "message": "Why is the sky blue?",
  "learner": { "age": 9, "language": "en" },
  "requestedDepth": "balanced"
}