A FelixChess Innovation

Move Tree

The best way to learn openings

About FelixChess

The place you go
before you play.

What FelixChess is

  • A platform to study, explore, and deeply understand chess openings
  • An interactive Move Tree - the modern answer to a question the chess world has been asking since 1981
  • A preparation tool: study here, then use what you learned on any playing platform
  • Built on millions of real games, making the data behind every opening line statistically meaningful

What FelixChess is not

  • Not a place to play games - for that, chess.com and lichess.org are outstanding
  • Not a live tournament viewer - dedicated broadcast tools do that better
  • Not a competitor to chess.com or lichess - FelixChess exists alongside them, not against them
"chess.com and lichess.org are fantastic experiences for playing and watching chess. FelixChess is where you go before - to study, prepare, and understand the openings you will play."

The gap FelixChess fills

In 2026, the most common way to learn chess openings is still watching videos, reading books, and navigating flat two-dimensional tables. The most advanced playing platforms in history co-exist with opening study methods that have not changed in decades. FelixChess is the modern answer: an interactive, data-driven Move Tree built from millions of real games, where every branch represents actual human decisions at the board.

games in the repository
Create Free Account

Technology

Built on foundations that
scale.

Aegis XP
A composable experience platform built on OSGi and Apache Sling - the same technology foundation used by enterprise content platforms worldwide. Every FelixChess component is independently deployable and the platform scales horizontally.
Apache Jackrabbit Oak
A content repository designed for billions of nodes. FelixChess stores every game, player, and opening position as a node in a virtual hierarchy - the same tree structure the original 1981 concept described, now running at scale.
AI integration
Stockfish 18 runs in-browser via WebAssembly for position analysis with no server round-trip. Natural language game search is powered by a large language model - ask for games in plain English and the platform finds them. Voice input is in development.
Scale
Designed for 30 million games and growing. The architecture makes no assumptions about an upper limit.