// the find
labuladong/fucking-algorithm
Crack LeetCode, not only how, but also why.
A Chinese-language algorithm study guide built around LeetCode problems, focused on teaching the underlying patterns rather than just posting solutions. The author's core argument is that most interview prep material gives you answers without mental models, and that a small number of reusable frameworks (sliding window, two pointers, DP state transitions, backtracking templates) covers most of what you'll face. 134k stars makes this one of the most-starred educational repos on GitHub.
The framework-first approach is genuine — rather than cataloguing 500 isolated solutions, articles explicitly teach the decision tree behind picking a technique. The table of contents is unusually well-organized: data structure fundamentals, then algorithm paradigms, then practice sets, with a clear dependency ordering that lets you work through it linearly. Visualization tooling (the algo panel mentioned in the README) is a real differentiator; animated step-throughs of recursion and data structure mutations are actually useful for the topics that trip people up. The breadth is real: KMP, segment trees, Union-Find, Dijkstra, and DP space optimization are all covered, not just easy array problems.
The repo itself is a thin index — almost all content lives behind the labuladong.online paywall, which is not disclosed upfront. The free GitHub content is essentially a marketing page for the subscription. English readers get a degraded experience; the English version exists but is secondary and some articles haven't been translated. The 'framework' selling point occasionally oversells: not every problem fits a template, and readers who internalize only the patterns can struggle when a problem requires genuine insight rather than template application. Last meaningful commit was February 2026, and the 'updating ing' placeholders suggest several sections were abandoned mid-write.