How Recommendations Work

How feeds, shops and playlists decide what you see — for product people, curious readers and engineers.

By Pavel Kordík, Eva Nečasová and the Recombee team

Examples from optional

I want to

Examples from many platforms · clear explanations first.

Sections take 2–5 minutes. Your progress stays on this device.

More ways in

🌱 Living book (recommended)
See tellings shared by other readers, and generate or remix your own. Always labelled; the verified core stays untouched.

Features

Gamification
XP, levels, badges and a completion certificate. Without it, the book is a plain reader.

Personalized recommendations
Powered by Recombee. Without it, content is shown in default order.

Spaced repetition
Recall quizzes so you actually remember. Without it, no review prompts.

Missions
Story-driven paths with a final boss quiz. Without it, the Missions tab is hidden.

Mini-games
Short interactive games inside sections. Without them, game blocks are hidden.

AI takeaways
Key points shown as side notes next to each section. Helps you remember the essentials.

Built on research from Recombee Labs & FIT CTU Prague, with AI dětem. More in the project README.

Supported by Recombee, AI dětem, Google.org and TACR

Open source on GitHub · Create your own p-book

Sister project: SlAIdy — slides that stay markdown & SVG, edited together with AI (GitHub)

🔍

Ask the book Answers come from the book’s own concept summaries — no AI model, no guessing.

Ask a question in your own words — for example “what is a filter bubble?”. I find the concept that answers it, give you the book’s short answer and the section to read. If the book does not cover it, I say so, and you can send the question to Pavel and Eva with “Message the authors”.

Settings

Appearance

Accent colour (more unlock as you read)

Font size

Features

Gamification (XP, levels, badges) Personalization (Recombee) Spaced repetition Missions Mini-games AI takeaways (side notes)

About & contact

By Pavel Kordík, Eva Nečasová and the Recombee team · GitHub · sister project SlAIdy

Your data