✦ Liang AI · Product Guide LoopKey PointsLesson AI Insights3D·FormulasExam Trends Real-World ApplicationBy StageLearning Traces Targeted SolutionsFor ParentsAll Features 🌐 English Enter System →
AI Learning System · International Education Vertical Agent v2

Master A-Level with a scientific closed-loop system,
built on syllabus points and real exam data.

Covering 5 major subjects, two exam boards (Edexcel + CIE) — integrating syllabus breakdown, in-depth lecture notes, Liang AI analysis, 3D formula demonstrations, and targeted exam practice. It organizes Learn · Practice · Test · Review into a trackable, quantifiable learning loop. This isn't a generic AI chat — it's rooted in your syllabus, your past papers, and your mark schemes.

0
Syllabus points
bullet-level precision
0
Past exam questions
with official mark schemes
0
Full past papers
Edexcel + CIE
0
Bilingual lecture notes
9-stage deep teaching
0
+ 3D interactive labs
adjust parameters, see instantly
0
Knowledge graph nodes
visualized connections
⊕ The above is verified, live, and ready for you to use directly — just the tip of the iceberg. Behind it lies an even larger origin corpus of question banks / textbooks / answer materials, where the Liang AI Education Model, combined with multiple LLM foundations and grounded in real syllabi, past papers and mark schemes, is trained and engineered into this vertical AI agent for international education.
📚 Syllabus coverage, point by point🤖 Liang AI vertical agent 📖 Millions of words of deep lecture notes🎯 Question ⇄ syllabus bidirectional loop 🧪 3D formula interactive lab🧠 302-node knowledge graph 📈 Data-driven exam trend analysis✍️ Timed exams · handwritten real papers 📅 21-day sprint📊 Learning trace · error attribution 🧮 Scientific calculator🔍 ⌘K instant site-wide search
Core Philosophy · The Scientific Learning Loop

Not a Sea of Questions, but a Closed Loop: Start from "I Don't Get It," Go Full Circle to "I Can Ace the Exam"

Acing A-Levels isn't about doing more questions—it's about making sure every action connects seamlessly. Stuck on a topic? That's where you start — Locate → Master → Clarify → Practice Past Papers → Reconnect to the Topic → Review & Sprint. Each step links to the next; miss one, and you've got a gap. Liang AI connects the entire loop.

🔄
Scientific Loop
Go full circle, truly understand
❓
Start
What you don't know
Pinpoint the topic
📚
Master
Deep-dive notes
9 segments + scoring tactics
🤖
Clarify
Ask Liang AI
Real-time reasoning
🧪
Build Intuition
3D · Formulas
Adjust parameters, visualize
🎯
Practice Past Papers
Do real exams
Check against mark scheme
🔁
Reconnect
Analyze the topic
Questions ⇄ Topics, two-way
📈
See Trends
Spot high-frequency
Data-driven focus
📅
Review
21-Day Sprint
Categorize mistakes, re-learn
📖
Just starting a unit or reviewing after class? Use this loop to anchor new knowledge to the syllabus: Start from the topic → Master the notes → Do examples → Ask Liang AI if stuck → Return to the topic to confirm you've got it. Don't just skim the textbook—read with the topic goal in mind.
🗺️
Open Syllabus
Locate unit + topic
→
📖
Deep-Dive Notes
9-segment lecture + scoring tactics
→
🔖
Bookmark + Notes
Highlight / mark difficult points
→
📝
In-Note Examples
Practice as you learn
→
🤖
Ask Liang AI
Concept reasoning / whiteboard
→
✅
Return to Topic
Confirm mastery → Next topic
1

Check the Syllabus First: Know What to Learn

Open the subject syllabus, see which topics are in this unit and how many past paper questions each one has. Understand the weight of this unit in the exam, and start with a goal.

🛠 Sidebar → Subject → Syllabus
2

Read the Notes: Not Skimming, but Deep-Dive Learning

The deep-dive notes follow a 9-segment teaching method, with scoring tactics, common traps, and Command-Word strategies in the right column. Bookmark key points, take notes on difficult ones.

🛠 Subject Homepage → Study & Read
3

Ask Liang AI + Verify

Stuck on something in the notes? Call up Liang AI from the bottom-right corner and ask directly. After learning, go back to the topic page and confirm "I understand this topic." Finish one topic before starting the next.

🛠 Bottom-right ✨ / Topic Detail Page
🎯
After finishing a unit, test your mastery with past papers. The core is tracing every wrong answer back to a topic—not "I got this question wrong," but "I haven't mastered this topic." Do a question, analyze it, fill one gap. No question is wasted.
🗂️
Select Past Papers
Filter by unit/topic
→
✍️
Timed Practice
Handwritten / MCQ / Formulas
→
✅
Check Answers
Against mark scheme
→
🎯
Analyze Topic
What does this question test?
→
❌
Categorize Mistakes
Wrong → Which topic?
→
🔁
Verify with Similar Questions
One more question → Confirm mastery
1

Select Questions: Filter by Topic or Unit

In the past paper bank, filter by unit/topic, or do a full paper. Each topic shows its hit count—practice high-frequency ones first.

🛠 Subject → Past Papers / Sidebar → Exam
2

Do + Analyze: Trace Every Question to a Topic

After answering, click "🎯 What does this test?" → See which topics are hit and related note chapters. Wrong answers go straight to your mistake book, categorized by specific topic.

🛠 Bottom of practice page → Analysis Trio
3

Fill Gaps + Verify with Similar Questions

After categorizing a mistake, click the topic to study the notes or ask Liang AI. Then click "🔁 One more question on this topic" — do another one to confirm you've truly got it. If not, re-learn and practice again.

🛠 Topic Detail → Notes → Liang AI
🔁
The mastery heatmap and mistake book in your learning progress show you exactly where your weak spots are. Prioritize red topics—go back to the notes, ask Liang AI, do similar questions, and keep at it until they turn green. Filling a gap isn't "reading it again"—it's "getting one more question right."
📊
Check Mastery
Heatmap · Red = Weak
→
❌
Open Mistake Book
Categorized by topic
→
📖
Re-study Notes
Deep-dive + scoring tactics
→
🤖
Ask Liang AI
Reasoning + whiteboard
→
🔁
Verify with Similar Questions
Get one more right
→
🟢
Mastery Updated
Red → Green
1

Heatmap Pinpoints Weak Spots

Open your learning progress. Red on the mastery heatmap = weak topic, spotted instantly. Your consecutive study streak shows how long you've kept at it. Data-driven, not gut-feel.

🛠 Sidebar → Learning Progress
2

Mistake Book: Fill by Topic

Mistakes aren't piled by time—they're categorized by topic. How many times you got P1.3 Quadratic Functions wrong, how many times P1.1 Exponents wrong—it's all clear. Prioritize the topics you've messed up the most.

🛠 Learning Progress → My Mistakes
3

Verify After Filling, Until It Turns Green

From the mistake book, click into the topic detail → study notes → ask Liang AI → do similar questions. Get it right, and your mastery updates automatically. One red topic turns green at a time. Fill one, secure one.

🛠 Topic Detail → Notes → Liang AI → Similar Questions
📅
21 days before the exam, follow the system's daily battle plan: one theme per day, with topic review + whiteboard formulas + matching past papers. Weak spots get priority (driven by your mistake book data), high-frequency topics get focused attack (driven by trend data). Not blindly flipping through books—prioritize with data.
📈
See Trends
Top 10 high-frequency topics
→
📅
21-Day Plan
One theme per day
→
📖
Daily Study
Topic + whiteboard + past papers
→
⚡
Weak Spots First
Driven by mistake book
→
🎓
Mock Exam
Timed full paper
→
🔁
Re-learn & Practice
Until the day before the exam
1

See Trends: Data Sets the Focus

Open the question trend analysis, see the Top 10 high-frequency topics from the last 3 years and their trends (Stable / Up / Down). Where the trend points, that's where your effort goes.

🛠 Subject → Trend Analysis
2

21-Day Sprint: Execute Day by Day

Each subject has a sprint plan, Day 1–21 with one theme per day. Each day includes: topic review + whiteboard formula recap + matching past papers + your weak spot alerts. Just follow it—no planning needed.

🛠 Sidebar → 21-Day Sprint
3

Sprint Review + Mock Exam

Sprint review quick-reference: command verbs + must-memorize formulas + glossary + exam key points recap. Do 1–2 timed mock exams before the real one, submit for AI analysis, then fill your weak spots one last time.

🛠 Subject → Sprint Review / Sidebar → Exam
Why Liang AI is Different

Generic AI forgets as soon as you finish chatting.
Liang AI is rooted in syllabi, past papers, and mark schemes — every step connects.

Generic AI chat tools can't handle A-Level precisely: they don't know the Edexcel vs CIE syllabus differences, they don't know how mark schemes award points, they don't know how this question has appeared in past exams. Liang AI has mastered all three.

❌ Generic AI Chat

  • Chat history doesn't stick — ask the same thing again next time
  • Unaware of mark scheme points: answer is "correct" but misses M1/A1
  • Can't link to specific topics — study for hours without knowing where you are in the syllabus
  • No past paper basis — may confidently produce nonsense
  • Won't tell you how many times this question has appeared or the trend
  • Math formulas render poorly, chemical structures unrecognizable
  • No tracking: where did you leave off yesterday? Which topics are weak?

✅ Liang AI Vertical Agent

  • Syllabus-level granularity: 3,425 topics, each linked to past papers, notes, and trends
  • Past paper driven: 19,828 exam questions + official answers + marking points — every answer is verifiable
  • Mark scheme annotated: M1 / A1 / B1 point allocation — learn "how to score," not just "how to solve"
  • Two-way closed loop: Questions ⇄ Topics navigate each other — get a question wrong → pinpoint the topic → master it → try another
  • Perfect board work: KaTeX real-time rendering — math, chemistry, physics formulas as clean as real exam papers
  • Full traceability: progress, mistakes, bookmarks, mastery heatmap — every step remembered
  • Exam trends: high-frequency topic ranking + last 3 years' trends — data tells you what to focus on
Around the Syllabus · Exhaustive Coverage

Every Topic is a Web Connecting Past Papers, Lessons & Trends

We break down 5 subjects into 132 units / 3,425 bullet-level topics — down to every syllabus line. Click any topic to see: how many past paper questions it hits, historical trends, the corresponding lesson chapter, and content confidence levels (verified / reference). That’s what we mean by "exhaustive coverage."

Topic Detail = Learning Hub

Dive in from "I don’t get it" and radiate outward

  • See at a glance a topic’s past paper hits and 3-year trend
  • «Read the full lesson» jumps to the lesson chapter for systematic study
  • Content labeled verified (two-way evidence) / reference / inferred — no shortcuts
  • Two-way loop between questions and topics: trace all past paper questions from any topic
Topic Detail
Topic List · Question Density

See at a glance which topics are tested most

  • Each topic shows past paper hit count (e.g., Quadratic graphs → 56 questions)
  • Sparkline mini trend chart — up or down over the last 3 years at a glance
  • Laid out by unit for exhaustive coverage — no topic left behind
  • Question-to-topic mapping built from real exam questions
Topic List
Core Feature · Penetrating Lecture Notes

Not a textbook copy, but A* deep-dive lectures

Each subject features millions of words of bilingual (Chinese-English) deep-dive content. Every unit is structured with our 9-Segment Teaching Method, embedding real exam questions, scoring techniques, and A* depth targets — this is our strongest content moat.

9-Segment Penetrating Structure

Learning Objectives → Real Exam Questions → Common Mistakes → Metacognition

  • ① Learning Objectives ② Prerequisite Review ③ Concept Deep-Dive ④ Concept Diagnosis ⑤ Exam Mapping
  • ⑥ Examples → Real Exam Questions ⑦ Error Diagnosis ⑧ Retrieval Practice ⑨ Metacognition + Cross-Chapter
  • Right sidebar: A* Depth Targets · Scoring Techniques · Common Traps · Command-Word Strategies
  • Seamless bilingual toggle, with bookmarks / notes / reading progress fully tracked
Penetrating Lecture Notes
Formula / Relationship Cards · Scoring Cues

Every formula tells you "when to use it in exams"

  • Formula Card = Formula + Applicable Boundaries + Exam Trigger Clues
  • e.g., $\Delta=b^2-4ac$ → "Use when the question asks about discriminant / number of roots"
  • Turning "knowing the formula" into "knowing how to score"
Formula Relationship Cards
Liang AI · Exam-Level Explanations

Ask once, Liang AI renders a full-mark solution in real time

Generic AI often "confidently makes things up"; Liang AI is grounded in real syllabi, past papers, and official mark schemes, rendering each step into proper mathematical notation, with M1 / A1 scoring points clearly marked. Below is a real example · live rendering (not a screenshot):

Full-Mark Solution · Real-Time Formula Rendering
Past paper (P1 2019 Jan Q1 adapted): Simplify $\dfrac{2}{\sqrt5-1}$, writing in the form $a+b\sqrt5$.
① Rationalize denominator (multiply by conjugate)M1
$\dfrac{2}{\sqrt5-1}=\dfrac{2(\sqrt5+1)}{(\sqrt5-1)(\sqrt5+1)}$
② Simplify denominator
$=\dfrac{2(\sqrt5+1)}{5-1}=\dfrac{2(\sqrt5+1)}{4}$
③ Expand and simplifyA1
$=\dfrac{1}{2}+\dfrac{1}{2}\sqrt5\quad\Rightarrow\quad a=\tfrac12,\ b=\tfrac12$
✓ Verified · Matches past paper · Aligned with mark scheme
🤖 Liang AI · International Education Vertical Agent
Liang AI real conversation
↑ Real conversation: ask "bubble sort exam focus", Liang AI explains working principle + step-by-step demo + high-frequency calculation questions — for any subject, any question, explained thoroughly using the syllabus and past papers.
Modeling · 3D · Formula Demo

Turn abstract formulas into intuition

Science Lab with 66+ adjustable sandboxes: drag parameters and watch curves change in real time — functions, mechanics, enzyme kinetics, titrations — all playable. Plus a 302-node knowledge graph + formula wall + math animations that weave knowledge into a visible web.

Quadratic $y=ax^2+bx+c$ · Drag a to see the opening
Sine wave $y=A\sin(\omega x+\varphi)$ · Real-time oscillation
302-Node Knowledge Graph

Your entire subject, woven into one visible network

  • Knowledge Graph · Mind Map · Formula Wall · Animations · Learning Pathways — five in one
  • Click any node → see core positioning, exam hooks, and top-scoring writing tips
  • Move from isolated memorization to connected understanding
Knowledge Graph
Science Lab Sandbox

Tweak parameters, build intuition — stronger than rote learning

  • Quadratic functions, sine waves, exponential decay, logistic growth, normal distribution…
  • Run / Refresh for real-time plots + Insight prompts (e.g., axis x = -b/2a)
Science Lab
Real Exam-Based · Intelligent

Precision effort on high-frequency priorities

1,881 past papers analyzed by topic for frequency and trends: which topics are high-frequency must-knows, which have risen over three years — data speaks. Then combine timed exams + real-paper handwriting to turn practice into real combat.

0
Past papers · Official answers + marking points
0
Years of exam trend data
🔥 Top 10
High-frequency topic ranking + Sparkline
3 Modes
Answer: Handwriting / MCQ / Formulas
Exam Trends · Top 10

Priorities aren't what teachers say — they're what exam data tells you

  • High-frequency topic ranking + last 3 years trend (Stable / Up) + Sparkline
  • Multi-line chart of years × unit coverage — hot and cold at a glance
Exam trends
Timed Exams · Real-Paper Handwriting

Just like the real thing: timed, page-turning, handwritten answers

  • Real-paper timer (44:35), question navigation, Flag for review
  • Draw / Handwrite / Input formulas on the real paper — answer exactly like in the exam
  • Submit then AI grading & topic analysis
Real-paper handwriting
Common Student Struggles × Liang AI Solutions

Every learning challenge has a matching solution in the system

Every A-Level student has faced these scenarios. Traditional methods mean toughing it out; Liang AI gives you the solution.

😵
Common Struggle

"I've done tons of practice, but my score won't budge."

It's not about doing too few—it's about not knowing where you went wrong. Liang AI automatically traces your mistakes back to specific exam topics. You're not getting "that question" wrong—you're getting that topic wrong. Go back and master the topic, then verify with similar questions.
See it in action → Error attribution + closed-loop review
📚
Common Struggle

"The textbook is too thick—I don't know what's important."

The syllabus is your map. Every topic is tagged with how many times it's appeared in past papers and its trend over the last 3 years—the high-frequency ones are your priorities. The guided notes condense the textbook into 9 sharp lectures, telling you exactly "how it's tested and how you should answer."
See it in action → Syllabus navigation + guided notes
🤖
Common Struggle

"I ask ChatGPT, forget the answer, and have to start over."

Liang AI's answers are rooted in your learning history—questions you've asked are linked to topics, topics to notes, notes to past papers. Next time you open the system, pick up exactly where you left off. No need to start from scratch.
See it in action → AI chat + learning history
😰
Common Struggle

"I have no idea what to review before the exam."

The 21-day sprint is organized day by day: one topic per day—key points review, board-work formulas, and matching past paper questions. Your error log and mastery heatmap show your weak spots, so you prioritize those. The Top 10 high-frequency topics are backed by real data.
See it in action → 21-day sprint + trend analysis
🧮
Common Struggle

"I memorized the formula, but I can't apply it in problems."

Formula cards in the notes tell you "when to use this in an exam"—e.g., $\Delta = b^2 - 4ac$ → "Use when the question asks about discriminant / number of roots." The 3D lab lets you drag parameters and see curves change in real time, turning abstract formulas into intuitive understanding.
See it in action → Formula cards + 3D Lab
🌐
Common Struggle

"Knowledge points feel scattered—I can't connect them."

A 302-node knowledge graph + mind map + formula wall + learning path weaves the entire subject into a visible network. Click any node to see its core definition, exam hook, and top-scoring answer style. Move from isolated memorization to connected understanding.
See it in action → Knowledge graph navigation
Practice-Driven · Real Materials, Real Results

One Core Thread: Quadratic Functions & the Discriminant, a Six-Step Closed Loop Built on Authentic Content

Every step below is real content from our system — real syllabus items, real lecture notes, Liang's real derivations, real full-mark solutions. No filler descriptions.

① Pinpoint the Exam Topic · Real Syllabus (bullet level)
📚 Edexcel International A-Level Mathematics · P1 Pure Mathematics 1
  • P1.1 Laws of indices for all rational exponents Laws of indices for all rational exponents
  • P1.2 Use and manipulation of surds Use and manipulation of surds
  • P1.3 Quadratic functions and their graphs Quadratic functions and their graphs 🔥 Top 3 High-Frequency
  • P1.10 Algebraic manipulation of polynomials Algebraic manipulation of polynomials
First, identify what you need to master per topic → lock onto P1.3 Quadratic Functions, a high-frequency must-know. Start here.
② Deep Dive into Lecture Notes · Quadratic Functions Board (Real Formulas)
General Form $y = ax^2+bx+c\quad(a\neq 0)$
Discriminant $\Delta=b^2-4ac$ 💡 Triggered when the question asks "number of roots"
Axis of Symmetry $x = -\dfrac{b}{2a}$
Vertex Form $y=a(x-h)^2+k$ (Vertex $(h,k)$)
Completing the Square $ax^2+bx+c=a\left(x+\frac{b}{2a}\right)^2+\left(c-\frac{b^2}{4a}\right)$
Multiplication $a^m \times a^n = a^{m+n}$
Fractional Indices $a^{\frac{1}{n}} = \sqrt[n]{a}\,,\quad a^{\frac{m}{n}} = \sqrt[n]{a^{m}}$
⭐ From the lecture notes: "The discriminant is the foundation that runs through P1–P4. Master the discriminant + completing the square, and quadratic functions are yours."
🔆 High-Frequency Pitfalls · Real Trap Warnings
  • Index Confusion $(a^m)^n = a^{mn}$ is not $a^m\times a^n$ —— $(x^2)^3=x^6$, not $x^5$.
  • Rationalizing Forgetting to multiply by the conjugate or only multiplying the numerator; for $\dfrac{1}{2+\sqrt{3}}$ you must multiply by $\dfrac{2-\sqrt{3}}{2-\sqrt{3}}$.
  • Completing the Square Forgetting to factor out the leading coefficient; for $2x^2+8x+5$ first write $2(x^2+4x)+5$ then complete the square.
  • Discriminant $b$ includes the sign! In $x^2-3x+1=0$, $b=-3$, $b^2=9$.
These "trap alerts" come from real lecture notes / 21-day content. Scan them before the exam to avoid easy point loss.
③ Ask Liang · Real Derivation (Real-Time Rendering)
🙋 Why can $b^2-4ac$ determine the number of real roots?
From the quadratic formula $x = \dfrac{-b \pm \sqrt{b^2-4ac}}{2a}$: the existence of real roots depends on whether $\sqrt{\;}$ is non-negative. If $b^2-4ac > 0$, $\sqrt{\;}$ yields a real number, and $\pm$ gives two distinct values → two distinct real roots (the curve intersects the $x$ axis at two points). If $=0$, the square root is $0$, and $\pm$ produces no new value → one repeated root (the curve touches the $x$ axis). If $<0$, the square root of a negative number is undefined in real numbers → no real roots (the curve lies entirely above or below the $x$ axis). So $\Delta = b^2-4ac$ is the "judge" of the relationship between a quadratic function and the $x$ axis.
④⑤ Solve Real Exam Questions + Analyze Back to Topics · Full-Mark Solutions (Real Exam Content)
Exam Question (P1 2019 Jan Q1 adapted): Simplify $\dfrac{2}{\sqrt{5}-1}$, writing it in the form $a+b\sqrt{5}$
① Rationalize the denominator $\dfrac{2}{\sqrt{5}-1}=\dfrac{2(\sqrt{5}+1)}{(\sqrt{5}-1)(\sqrt{5}+1)}$ M1
② Simplify the denominator $=\dfrac{2(\sqrt{5}+1)}{4}$ A1
③ Expand to get the answer $=\dfrac{1}{2}+\dfrac{1}{2}\sqrt{5}\ \Rightarrow\ a=\tfrac{1}{2},\ b=\tfrac{1}{2}$ A1
Full-mark solution board
After solving, tap 🎯What topic does this test? → 📚 Hits P1.2 Surds / P1.1 Indices → 🔁 Try another question on the same topic — a closed loop between questions and syllabus topics.
🧪 Build Intuition · 3D / Formula Demo (Science Lab)
Quadratic Function $y = ax^2 + bx + c$ · Drag parameter $a$ to see the direction of opening
Real-time hint: $a>0$ opens upward, $a<0$ opens downward; axis of symmetry $x = -\dfrac{b}{2a}$
Science lab
66+ adjustable sandbox: drag parameters to see the curve change in real time. Turn "discriminant / vertex / opening direction" into intuition — more effective than rote memorization.
⑥ 21-Day Sprint · Day 1 Real Schedule
📅 Day 1 · Algebra & Functions Basics
Today's syllabus topics: P1.1 Laws of Indices · P1.2 Surds · P1.3 Quadratic Functions · P1.10 Polynomials — supported by board formulas + exam practice. Prioritize weak areas and conquer P1–P4 in 21 days.

🧬 The same applies to Biology: T1 Cell Structure → §1.1 Microscopy (lecture board $\text{Magnification}=\dfrac{\text{image size}}{\text{actual size}}$) → Ask Liang "What's the resolution difference between light and electron microscopes?" → T1 exam questions (including microscope calculations) → Analyze back to §1.1/§1.2 → 21-Day Day 1 is exactly Cell Structure. The same closed loop works for any subject.

By learning stage · not isolated features, but one learning thread

Each stage, exactly how to use it — it only works when strung together

There are many features, but for one student they form a line that unfolds with the learning process. Below, by real stage: what you're doing → which features → how it connects to the next step. Stages interleave and run in parallel, but all hang on one "syllabus topics" backbone.

📖
In class · same day / after class

Master each lesson's topics the same day

Scenario: you didn't fully follow the lesson, or want to consolidate the chapter after class.
  • ① Use the syllabus to locate this chapter's topics, see which are high-frequency
  • ② Penetrative lecture: 9-part teaching + formula cards tagged "when it's tested"
  • ③ Ask Liang AI about anything unclear (grounded in your topics, not generic)
  • ④ Save class notes to Study Archive · Notes (Markdown + formulas + live preview)
🔗 Notes link to topics, topics to lectures — later practice & sprint pull straight from here, no re-learning.
Syllabus topicsPenetrative lecturesAI chatArchive·Notes
🎯
Topic deep-dive · a topic you keep missing

Learn one topic until it's intuition

Scenario: a topic you keep getting wrong, or an abstract formula you can't apply.
  • ① Syllabus shows its past-paper hits + 3-year trend — decide how much to invest
  • ② Lecture deep-dive + high-frequency pitfall warnings
  • ③ 3D Science Lab: drag parameters, watch curves change live — turn formulas into intuition
  • ④ Verify with same-topic past papers; real mistakes go to the error book
🔗 Master one node and the knowledge graph links it to neighbors — from isolated memory to connected understanding.
Syllabus trendsLectures/formula cards3D LabKnowledge graph
✏️
Practice · question⇄topic

Practice with purpose — every question maps to a topic

Scenario: doing past papers, looking up a source question, or drilling by topic.
  • ① Source-question lookup: image / text / PDF → instant top-K + confidence
  • ② Do past papers (full / selected / whole-paper, three formats)
  • ③ After answering, click "🎯 What does this test?" → hit topics → "🔁 One more on this topic"
🔗 Wrong answers auto-collect & attribute to topics — practice output feeds review and sprint.
Source lookupPast papersQuestion·topic loopAuto error book
🔁
Mistake review · online + offline

Online & paper mistakes — unified attribution + AI breakdown

Scenario: in-system mistakes, plus paper mistakes from school / tutoring.
  • ① System mistakes auto-attributed to topics, weak areas at a glance
  • ② Snap paper mistakes → upload to Study Archive · Mistakes
  • ③ Select → AI mistake analysis: recognize the question + core topic + key takeaways + thorough analysis + detailed worked solution + weak-point summary
  • ④ Annotate on the image; the report auto-saves for later review
🔗 Every mistake (online + paper) lands in one place, attributed to one topic system — evidence-based review, targeted gap-filling.
Error attributionArchive·MistakesAI analysisAnnotation
🌙
Daily after-class · custom deep-dive

Each day's key/hard points, AI-analyzed against the syllabus

Scenario: after class each day, you want to systematically review the day's key / difficult points.
  • ① Custom Study: pick the day's topics / difficulties; AI generates a key/hard-point analysis grounded in syllabus + lectures + past papers
  • ② Ask Liang AI on stuck questions for live derivations
  • ③ Save output to Study Archive · Notes
🔗 A little every day, all saved to the archive — reused directly at finals / sprint, no duplicate work.
Custom StudyAI chatArchive·Notes
🏁
Chapter / subject end · system review

Finish a round — first see where you're weak

Scenario: a chapter or subject wraps up; you want to assess mastery.
  • ① Mastery heatmap: which topics are weak, by color
  • ② Trends: see the high-frequency Top topics
  • ③ Knowledge graph full review + mistake summary
🔗 Turn "finished" into "see the weak spots" — directly produces the next step's (sprint) priorities.
Mastery heatmapTrendsKnowledge graphMistake summary
🚀
Pre-exam sprint · final stage

Converge a semester's work into one sprint line

Scenario: N days before the exam, limited time — put effort where it counts.
  • ① 21-Day Sprint, day by day: topic recap + board formulas + matching past papers
  • ② Use the error book + mastery heatmap to prioritize; weak points first
  • ③ High-frequency Top 10 — data picks the focus
  • ④ Online mock exam: timed, page-turning, handwritten answers — real rehearsal + AI mistake analysis to plug gaps
🔗 A whole semester's notes / mistakes / mastery converge here into the final sprint route.
21-Day SprintError/heatmapHigh-freq Top10Online mock

🧵 One thread: class notes → topics you deep-dive → mistakes from practice → daily custom analysis → end-of-course mastery → sprint priorities — all saved in the Study Archive, attributed to one topic system. That's the "thread" — not a pile of isolated features.

Learning Trail · Pick Up Where You Left Off

Every Step You Take, the System Remembers

Progress is clearly visible — one click to resume where you left off. Wrong answers are automatically collected and traced back to specific topics. Bookmarks, tags, and text notes can be added anytime — making review data-driven and reliable.

Personal Space · Learning Trail

Progress · Mistakes · Tags · Notes — All Tracked

  • Auto-save position: Whether you're reading a handout or solving problems, come back and pick up where you left off with one click.
  • Error log: Wrong answers are automatically collected and traced back to specific topics, so you review exactly what needs work.
  • Bookmarks & Tags: Mark key sections with one tap and revisit anytime.
  • Text notes: Take notes as you learn, attached to specific topics or handout sections.
  • Streak tracking + Mastery heatmap: See your consistency at a glance and pinpoint weak spots instantly.
Learning progress and error log
For Parents · See How Your Child Learns

All that money spent on tutoring,
is it actually working? Now you can see for yourself.

Traditional tutoring: your child goes to class → you have no idea what they learned → you only see the result on exam day — too late. Liang AI makes the learning process fully visible, trackable, and measurable from start to finish.

Real-Time
Mastery heatmap
See which topics are strong
and which need work at a glance
Automatic
Error analysis by topic
Not "I got this question wrong"
but "I haven't mastered this topic"
21-Day
Sprint plan visible
See what's studied and practiced daily
Parents can follow along too
Full Transparency
AI chat logs accessible
Every question asked and answered
Learning trail is fully traceable
👨‍👩‍👧

How can parents check?

Open the "Learning Progress" page — the heatmap shows your child's mastery level for every topic. Red = weak, green = mastered. You don't need to know the subject, just the colors.

🎯

How do we avoid wasting time?

Next to each topic, you'll see the number of past exam questions and a trend arrow. High-frequency, must-know topics get studied and practiced first. The system uses data to set priorities — no more guessing "what to study."

💰

How is this better than traditional tutoring?

One-on-one tutoring costs £40–80 per hour, and tutors may not be fully familiar with the Edexcel / CIE syllabus. Liang AI covers the entire syllabus + 19,828 past exam questions — at a fraction of the cost.

Product Moat · Uncopyable

Why can’t others replicate this?
It’s not a tech barrier — it’s content depth.

3,425
Syllabus points
deconstructed to bullet-level precision
Human + AI cross-verified
19,828
Past exam question links
Question → topic mapping
Bidirectional navigation
12.5 million words
Bilingual lecture notes
9-stage deep-teaching method
Scoring tactics + A* target
302 nodes
Knowledge graph
1,497 edges connecting them
Visualized pathways

Generic AI companies can call large language model APIs, but they cannot replicate the 3,425 deconstructed syllabus points, the 19,828 question-to-topic mappings, the 12.5 million words of deep-teaching notes, or the 302-node knowledge graph. That is Liang AI’s moat — deep content × vertical intelligence.

All Capabilities · At a Glance

See Everything: Liang AI's Full Capabilities

From learning to exams, from single points to connected threads — here's the complete capability map. Each card shows where to access it, so you can start using it right away.

📚 Learn · Syllabus & Lectures
📚

Syllabus · Full Coverage

5 subjects broken into 132 units / 3425 bullet-level topics. First, know exactly what you need to master.

Access: Sidebar → Subject → Syllabus
📖

Deep-Dive Lectures

Millions of words of bilingual (Chinese/English) content. Each unit uses a 9-step teaching method, with real exam questions, scoring tips, and A* targets embedded.

Access: Subject Home → Lectures
🔆

Tips Lightbulb

Key reminders and high-frequency pitfalls placed next to lecture paragraphs. Scan them before exams to avoid losing easy marks.

Access: 💡 markers inside lectures
🔖

Bookmarks · Tags · Notes

One-click bookmark key paragraphs. Attach text notes to topics or lectures. Keep track of what you've read.

Access: Inside lectures / questions
🤖 Ask · AI Vertical Agents
🤖

Liang AI Assistant

Your personal tutor, grounded in real syllabi and past papers. Renders formulas in real-time. Summon it from any page with a floating window.

Access: Bottom-right ✨ / Top bar AI
🎯

Question Analysis Trio

🎯 What this question tests · 📚 Which topics it hits · 🔁 One more question on the same topic. All marked "Verified · N past papers".

Access: Bottom of practice page
🧠

AI Mind Map · Knowledge Graph

302-node knowledge graph + mind maps + formula wall + animations + learning paths. Connect the dots.

Access: Sidebar → AI Mind Map
🧪

Science Lab

66+ adjustable sandboxes: drag parameters to see 3D models / curves change in real-time. Turn formulas into intuition.

Access: Sidebar → Science Lab
🎯 Practice · Test · Based on Real Papers
🗂️

Past Paper Bank

1,881 sets of past papers / 19,828 questions. Each with official answers and mark scheme breakdowns.

Access: Subject → Past Papers
✍️

Exam Mode

Timed, full mock exams. Answer by handwriting, MCQ, or formula input. Submit to get AI-powered analysis.

Access: Sidebar → Exam
📈

Question Trends

Top 10 high-frequency topics + 3-year trends + historical coverage by unit. Let data guide your focus.

Access: Subject → Trends
🔁

Topic Details (Hub)

For each topic: number of past paper hits, annual trends, linked lectures, and material confidence level.

Access: Subject → Topic → Details
📅 Review · Sprint · Planning
📅

21-Day Sprint

A battle plan per subject: one theme per day — topics, board formulas, and matching past paper questions.

Access: Sidebar → 21-Day Sprint
⚡

Sprint Review Quick Reference

Topic × textbook × key exam points recap. Includes command verbs, must-memorize formulas, and glossary.

Access: Subject → Sprint Review
🪪

Personal Space · Progress

Mastery heatmap + consecutive check-in streak + one-click return to where you left off.

Access: Sidebar → Personal Space
❌

Error Book

Wrong answers auto-collected and attributed to specific topics. Review and fix that one weak point.

Access: Personal Space → Errors
📂

Study Archive

Personal file space: study materials / notes (Markdown + formulas) / mistakes. Snap paper mistakes → AI mistake analysis (core points · key takeaways · detailed worked solution) + image annotation + inline PDF preview.

Access: Sidebar → Study Archive
🧰 General Tools
🔍

Site Search ⌘K

Cmd/Ctrl + K to instantly search lectures, past papers, or topics. Supports combined topic search across papers.

Access: Top bar search / Subject → Search
🌗

Bilingual + Theme

One-click switch between Chinese / English. Light/dark theme. Terminology naturally embedded in both languages.

Access: Top bar 中/EN · ☀️
🧮

Scientific Calculator

fx-991EX level: differentiation/integration/Σ/Π/SOLVE/matrix/statistical regression/natural 2D display.

Access: Top bar 🧮 button
💎

Account & Credits

Personal account and credit wallet. AI usage clearly visible.

Access: Sidebar → Account

Ready to close the loop?

Pick a subject, start from one topic, and let Liang AI guide you through the entire journey — wherever you're stuck, start right there.