Sitor

Sitor — AI 1-on-1 Tutoring. Learn Until You Truly Understand.

Sitor — AI 1-on-1 Tutoring. Learn Until You Truly Understand.

In 1984, education researcher Benjamin Bloom discovered that students who received 1-on-1 tutoring outperformed 98% of students in traditional classrooms. This became known as Bloom's 2-Sigma Problem — 1-on-1 instruction is extraordinarily effective, but far too expensive to scale.

Forty years later, AI changes the equation.


Why We Built Sitor

Have you ever had this experience: you watched a bunch of tutorials, felt like you "got it," then the moment you tried to apply it — your mind went blank?

That's not your fault. It's a flaw in how learning works.

The vast majority of tutorials, courses, and AI tools follow the same pattern — one-way information delivery. Auto-generate a mind map, and you think "wow, so clear" — then close the tab and remember nothing. Auto-summarize a book, and sure, you saved time — but that's the AI's understanding, not yours.

Psychology has a term for this: "fluency illusion" — because the information is presented so clearly, you mistake it for your own understanding. Real understanding isn't "I heard it." It's "I can explain it."

This is the problem Sitor exists to solve.


What Is Sitor

Sitor is an AI-powered 1-on-1 tutoring platform. Tell it what you want to learn. It assesses your level, generates a personalized learning roadmap, then uses Socratic questioning to guide you through each concept until you truly understand it.

No passive video watching. No flashcard grinding. Just you, thinking things through — with an AI that knows exactly where to push.

This is the fundamental difference between Sitor and other AI learning tools: we don't generate content for you to consume. We guide you to think. You're not a passive receiver — you're the one doing the learning.

Any topic. Programming, math, physics, economics, a book, a new field — if it can be broken down into concept nodes, Sitor can teach it.


Core Experience

🎯 Assess First, Then Customize

Everyone starts from a different place. Jumping straight to Chapter 1 wastes time for experienced learners and overwhelms beginners.

Sitor doesn't start by lecturing. It asks a few open-ended questions to gauge your actual level — not testing whether you've memorized definitions, but probing how deeply you understand the underlying concepts. Then it generates a learning roadmap tailored entirely to you, skipping what you already know and starting where you actually need.

Assess before teaching
Sitor is not checking memorized definitions. It is probing how you reason.
A.After every render unless React can prove the dependencies did not change.
B.Only after the first render, unless you call it manually.
C.Whenever state changes anywhere in the application.
Other (type your own)

💬 Socratic Questioning

When you answer a question, Sitor doesn't just say "right" or "wrong." It assesses whether you truly understand or are just surface-level.

If you're vague, it keeps probing. If you're wrong, it doesn't correct you directly — instead, it follows your logic to its conclusion, letting you discover the contradiction yourself. Only then does it pose the key question that guides you to rethink.

This is why Sitor's teaching feels fundamentally different from "asking ChatGPT a question and reading the answer" — every concept is something you figured out yourself, not something you were told.

Socratic questioning showcase

📊 Master It Before Moving On

Every concept has a real-time mastery score across four dimensions: accuracy, depth of understanding, novel application, and discrimination between similar concepts. You need 80% or above to unlock the next node.

No one can fake their way through. If the score isn't high enough, you keep working with the AI tutor until you truly get it.

Master before unlock
Effect lifecyclesMastered
Next node unlocked: Custom hooks
Effect lifecyclesMastered (84%)Unlocked: Custom hooks

🧠 Your AI Tutor Remembers You

Sitor automatically builds your learner profile during teaching — your background, strengths, common mistakes, and learning preferences. These profiles persist across courses. Next time you study something new, the AI isn't a stranger — it's a tutor who already knows you.

We've also built in SM-2 spaced repetition, scheduling reviews along the forgetting curve — ensuring you don't just "learn it" but actually retain it.

You can view and manage your profile under "My Profile" in the sidebar. It's your data, and you have full control.

🎨 Hand-Drawn Diagrams

When text isn't enough for abstract concepts, your AI tutor proactively draws diagrams — flowcharts, state diagrams, concept maps — in a warm, hand-sketched style.

This isn't passive viewing. The AI draws first, then asks you questions about the diagram, guiding you to understand relationships visually. Supports fullscreen view and picture-in-picture mode.

Hand-drawn diagrams
Closure = function + surrounding lexical scope

🎧 Language Learning

When you're learning a language, Sitor automatically activates listening and speaking exercises. Listening uses a multi-speed loop method — the same material at 7 speeds from slow to fast, building true comprehension rather than just exposure. Speaking offers record-and-score with word-level pronunciation feedback. These capabilities are built in — the AI weaves them naturally into the teaching flow.

Listening loops
Listening loops
Demo only. This block uses mocked progress and will not call audio APIs.
linking soundsweak formsmeaning before dictation
1
2
3
2x
5
6
7
Round 4 / 7 · 2x loop

Speed rises first, then drops back so you hear what you missed.

1x
“Could you walk me through the tradeoff before we commit to the change?”
3/7 rounds
Speaking feedback
Speaking feedback
/th//r//k.tʃ/
The architecture needs a cleaner boundary.
Repeat it, get scored word by word, then focus on the weak sounds.
Tap-to-record is intentionally disabled in the release doc.
Showing a sample pronunciation analysis state with mocked data.
Overall score91%
The65architecture88needs93a84cleaner79boundary95
Demo only. Recording and scoring are mocked here so readers can preview the real UI safely.

What's Next

What we're working on:

  • Mobile optimization — Learn anywhere, anytime. Full keyboard handling, touch interactions, and responsive design for phones and tablets.
  • Custom knowledge bases — Upload your own documents, notes, and textbooks. Let the AI teach from your materials, creating a truly personalized tutor built on your resources.
  • Deep research capability — Before teaching, the AI conducts in-depth research on the topic — synthesizing the latest papers, documentation, and industry practices — ensuring every session is grounded in professional, real-time information.

Final Thoughts

There's no shortage of AI tools. Ones that summarize articles, generate flashcards, search for answers — they're all great, but they solve the problem of information access efficiency.

Sitor tackles a different question: do you actually understand it?

We believe real learning isn't about seeing more information — it's about thinking things through yourself. AI's greatest value isn't thinking for you. It's guiding you to think.

Bloom posed the 2-Sigma Problem in 1984. It puzzled educators for forty years. Now, with AI capabilities where they are, it's time for a new answer.

If that resonates with you, give Sitor a try.

Start learning →

Invite a friend — you both get 30 credits.