the ai tutor that lives on your screen.
Clicky floats next to your cursor, sees everything you see, and teaches you anything — step by step, out loud, pointing at the exact pixel. No tab switching. No copy-pasting. Just ask.
works with Claude · GPT-4o · GitHub Copilot · Gemini · Ollama
how it works
no setup. no switching tabs. just ask.
Clicky is the first AI tutor built for how you actually work — always on, screen-aware, and right next to your cursor.
it floats next to your cursor
A tiny blue buddy lives 35px to the right of your cursor at all times. No window switching, no alt-tab. It's just there.
it sees your entire screen
Every time you ask something, Clicky captures a screenshot of all your monitors. It knows exactly what you're looking at.
hold a key and ask
Hold Ctrl+Alt+Space and speak. Or just say 'Clicky'. It hears you, processes your screen, and responds — all in under 2 seconds.
it points, explains, and teaches
Clicky flies to the exact UI element, draws a highlight ring around it, and explains step by step — like a real teacher beside you.
features
not just an assistant. a real tutor.
clicky.so built the concept. We built it for Windows — and added everything a real teaching tool needs.
Pixel-Perfect Pointing
Clicky uses Claude Computer Use to find the exact pixel of any UI element — on any DPI, any monitor. It flies to it and draws a highlight ring while it explains.
Knowledge Journal + Spaced Repetition
Every conversation is logged to a local SQLite database. Clicky reminds you to review topics at the right time using the SM-2 algorithm. Ask 'what should I review?' to start a flashcard session.
Real Tutor Behaviour
Step-by-step lessons that wait for you. Say 'next' to advance. Say 'repeat' to hear it again. Clicky never rushes you.
Whiteboard Annotations
Clicky draws arrows, circles, underlines, and labels directly on your screen as it teaches. Annotations fade automatically.
Wake Word — Always Ready
Just say 'Clicky' from anywhere. No hotkey needed. The mic is always listening locally — nothing is sent until you speak.
Document Context
Drag a PDF, DOCX, or TXT onto the panel. Clicky reads it and uses it as context for your next questions. Great for studying textbooks.
Lesson Recording
Record your screen as you learn — Clicky outputs an MP4 + a Markdown transcript of the full Q&A session.
Multilingual Auto-Detect
Speak in any of 15 languages. Clicky detects it automatically, responds in kind, and switches the TTS voice to match.
OCR Fallback
Say 'read the fine print' and Clicky runs Tesseract OCR on the screenshot — extracting text too small for vision models to catch.
Code Mode (Auto)
Opens VS Code, Cursor, IntelliJ? Clicky auto-detects and switches to a code-specialist mode — code blocks, language tags, algorithm breakdowns.
Workflow Capture
Clicky records your mouse clicks and keystrokes. Stop it and ask 'what did I just do?' — it narrates your workflow step by step.
Skills System
Drop a .py file in ~/.clicky/skills/ and create your own voice commands. No pull request needed. Infinite extensibility.
works with every ai
your keys. your models. your choice.
Switch providers from the tray in 1 second. No restart.
No keys? Use Ollama — runs 100% on your machine, completely free.
get clicky. it's free.
Works on Windows 10 and 11. No account, no subscription, no API key required to start. Runs locally with Ollama.