StudyKaKeyda

Guide

How to use StudyKaKeyda

StudyKaKeyda turns a question into a visual: a moving, step-by-step picture you can play, pause, step back through and change. This page follows one question from the empty box to a link you can send, with a picture of each stage. Most of it needs no account; history and sharing are the two parts that do.

1

Ask

Open the question box and choose how you want to ask: in words, as pasted code or as input → output examples. Fill in one box and press Visualize.

  • Three tabs sit above the box: Ask in words, Paste code and Input → output.
  • The words tab takes up to 2,000 characters. Try “a ball thrown at 20 m/s at 30°”, “3/4 + 1/6” or “reverse a linked list”.
  • Inside the box there is a camera button and a microphone button. Use them to photograph a printed question or to say it out loud.
  • No account is needed to ask. The question still leaves your browser: it goes to our server and, unless the built-in engine or the shared cache already has the answer, on to an outside AI provider (Google’s Gemini API by default). A tidied copy of it is kept in a shared cache that is not tied to you. The privacy page at /privacy lists what is kept, who receives it and for how long.
2

Watch

An engine works out each step and draws it: bars swap, arrows slide, tables fill, a ball flies. The player shows the picture, the code, the values, the explanation and the controls together.

  • Every panel follows one step number, so the highlighted code line, the values and the explanation always describe the same moment.
  • Where an engine covers the topic, every number is computed by that engine rather than guessed.
  • AI reads your question to pick the engine and its inputs, writes the code when you have not given any, and draws the visual itself only when no engine covers the topic.
  • When no engine covers a topic yet, the visual carries an “AI-generated” badge. Its tooltip says the numbers may need checking. “Report a wrong visual” is on the page either way.
3

Control

Play, pause, step back, jump to the next pass of a loop or the next call, and repeat a part. The buttons and the keyboard move the same playhead.

  • Space plays and pauses. ← and → move one step at a time.
  • Shift + → jumps to the next pass of a loop. Alt + → steps over a line of code.
  • Click a line number to set a breakpoint, a place where the player always stops. Give it a condition like i == 3 if you want it to stop only sometimes.
  • L sets a loop: press it once for the start (A), again for the end (B), and a third time to clear it.
4

Change the numbers

Open “Try your own values” and move a slider or edit a list. The engine works everything out again and keeps your place.

  • A number gets a slider with its current value beside the label. A list is a text box: write numbers separated by commas, like 5, 2, 9, 1.
  • The change is applied a moment after you stop moving, and the panel says “Rebuilding” while the picture is redrawn.
  • The redrawn steps come from the same engine that made the first run, so those numbers are computed rather than guessed.
5

Ask the AI

Stop on the step that confuses you and use “Ask about this step”. The answer is about that step, using its explanation and values as context.

  • Suggested openers include “Why does this step happen?” and “What changed here?”.
  • Answers come in English, Hindi or Hinglish, chosen in the player. The menus and buttons stay in English.
  • Your question, the step’s explanation, its values and part of your code are sent to our server and on to an outside AI provider (Google’s Gemini API by default) to produce the answer.
  • Signed in, the conversation is kept with the visual until you delete your account. Signed out, nothing is kept on our servers, and only your last three questions stay in this browser tab.
6

Keep it

Signed in, every question you ask is saved and reopens at the exact step you left. A share link is public and read-only, and it opens at the step you had last watched.

  • Signed in, every question you ask is saved on its own and reopens at the exact step you left. Ready-made Library visuals are not added to your history.
  • The History page lists everything you have visualized and lets you delete any question from it. Settings → Data & privacy is where you delete the whole account.
  • A share link is public: anyone who has it can open it. It starts at the step you had last watched when you made the link, and the viewer can move freely through every step from there. Rotate it to stop the old link working, or turn sharing off.
  • E saves the current step as a picture (PNG), and N opens study notes to print or copy.
  • Deleting one question takes it out of your history; the row is marked as deleted rather than erased. Deleting your account removes your history, saved visuals and AI conversations, but not the anonymous copy of the question held in our shared cache or published in the Library. The privacy page explains that.

Three ways to ask

The tabs above the question box. Pick whichever matches what you already have.

Ask in words

When: You have a question in plain language, or a photo of one from a book or a board.

How: Type it into the box: up to 2,000 characters of maths, physics, chemistry, biology or a coding problem. The camera button reads a photo of a printed question (JPG, PNG, WebP or HEIC, under 20 MB). Your browser shrinks it and sends it to our server, our server passes it to an AI provider that reads images (Google’s Gemini API, or OpenAI’s where the operator has added it), and the text drops into the box for you to check. The image itself is not stored on our side. The microphone button takes dictation in English or Hindi in the browsers that support it. We never receive the audio, because your browser does the transcribing; in most browsers that means the audio goes to the browser vendor’s own speech service, which is outside our control. Press Visualize, or Ctrl / ⌘ + Enter.

Paste code

When: You already have code and you want to watch it run, line by line.

How: Choose the language, paste up to 20,000 characters, then press Visualize. The code goes to our server either way. Python, JavaScript, TypeScript and PHP are only checked there and handed straight back (not stored, not logged, not sent to AI), then traced inside your own browser. Java, C, C++, Go and Ruby are traced on a separate self-hosted runner machine when it is switched on for the server, so that source leaves our API too: the runner writes it to a temporary folder and deletes it when the run finishes. Two optional fields help. “Call to trace” names the exact call to follow, such as bubble_sort([5, 2, 9, 1]); leave it empty if your code already prints or calls the function. “Input for the program” supplies what your program reads from standard input.

Input → output

When: You know what should go in and what should come out, but not how to get from one to the other.

How: Add between one and ten example rows, each up to 1,000 characters a side. StudyKaKeyda reads the rows, suggests a problem statement and asks “Is this the problem?”. Edit it if anything is off, then confirm. Each row becomes a test the solution must pass, and AI-written code is run against those tests in your browser before you see it. If a round of testing cannot confirm it, the visual is labelled “Couldn’t verify — showing best attempt”.

Getting more out of it

Your draft waits for you

An unfinished question is kept in your browser for seven days, one slot per tab of the question box, and offered back as “Restore draft?” when you come back. Sending the question clears that slot, and signing out clears them all.

The keys pause while you type

Player shortcuts are ignored whenever the cursor is inside a text field, so typing a question never triggers them. Click the visual once to hand the keys back.

Name the call you want traced

Leave “Call to trace” empty if your code already prints or calls the function. Fill it in, for example bubble_sort([5, 2, 9, 1]), when the file only defines things and never runs them.

One step, or a whole pass

← and → move a single step. Shift + ← and Shift + → jump a whole pass of a loop, which is much faster on a long loop.

See it in 3D

A visual with a solid shape, 3D vectors, a molecule, a crystal cell or a field gets a 2D | 3D switch in the player. In 3D, drag the picture to turn it, and pinch or use Ctrl / ⌘ + wheel to zoom; press D to switch between the two. Settings → Appearance → Visuals sets which one a visual opens in, and a device that cannot draw 3D shows the 2D drawing instead.

A photo is read once

The camera button sends the photo to our server, which passes it to an AI provider that reads images (Google’s Gemini API, or OpenAI’s where the operator has added it). Nothing is kept afterwards. Reading a photo needs the AI helper and counts against the daily AI limit.

Switching the code language uses AI

The player can show the same solution in another language. Doing that sends your code to an outside AI provider (Google’s Gemini API by default) to be translated. The result says it was translated by AI, so check it against your own code. Undo puts the original back.

What still works without a connection

The visual you had open keeps working, along with pages you have already visited. Asking a new question, reading a photo, Ask AI and syncing your history all need a connection.

A share link is open

Anyone who has the link can open the visual, read-only. It starts at the step you had last watched when you made the link, and they can move through every step from there. Rotate the link to stop the old one working, or turn sharing off.

For teachers and parents

Put a visual on a big screen and press P for presentation mode and F for full screen. Q turns on “Quiz me”, which asks the class to predict each step before it plays. Sign in to make a share link that opens at the step you choose, so students can replay it at home.

Keyboard

Inside the player. Press ? to see this list there.

Space
Play or pause
→
Step forward
←
Step back
Shift + →
Next iteration
Shift + ←
Previous iteration
Alt + →
Step over (code)
Alt + ↓
Step into (code)
Alt + ↑
Step out (code)
Home
First step
End
Last step
[
Slower
]
Faster
B
Toggle a breakpoint on the current line
L
Loop a section: set A, then B, then clear
Q
Quiz me: predict each step before it plays
S
Read each step aloud
F
Fullscreen
P
Presentation mode (hide the side panels)
E
Export this step as a PNG
N
Study notes (print or copy)
D
Switch between 2D and 3D (visuals that have both)
R
Re-centre the view
?
Show these shortcuts
1 – 4
Answer a quiz question; Enter continues
Ctrl + wheel
Zoom the visual (drag to pan)
← → ↑ ↓
Turn a 3D picture, once it has focus (Shift + arrows move it; + and − zoom; double-click to reset)
← / →
Resize the visual, once the divider beside it has focus (double-click the divider to reset)
Ctrl / ⌘ + K
Open search: find a Library visual or jump to a page
Ctrl / ⌘ + Enter
Send the question from the question box