An intelligent, high-performance tool for comparing legal documents and analyzing structural changes. Unlike generic diff tools, Law Compare understands the hierarchy of laws (Chapters, Sections, Articles, Clauses) and uses NLP to align content even when article numbers change.
🌐 Live Demo - Try it out without installation!
- Structural Diff: Automatically identifies chapters, sections, and articles. It detects added, deleted, modified, and even merged/split articles.
- Intelligent Alignment: Uses weighted similarity algorithms (Jaccard + Containment + Semantic) to track moved or renumbered articles.
- Multiple Views:
- Structural View: Tree-based visualization of changes across the entire document.
- Side-by-Side View: Dual-pane comparison for focused reading.
- Git-style View: Classic line-by-line diff for technical review.
- High Performance: Backend written in Rust with parallel processing and memory-efficient
Arc<str>data sharing. - NER (Named Entity Recognition): Automatically detects dates, amounts, and legal terms within changes.
| Structural Analysis | Side-by-Side Comparison |
|---|---|
 |
 |
| Git-style Line Diff |
|---|
 |
The easiest way to deploy Law Compare is using Docker with the included Caddy configuration.
# Clone the repository
git clone https://github.com/LikeRainDay/law_diff.git
cd law-compare
# Build and run with Docker
docker build -t law-compare .
docker run -d -p 8080:80 law-compareAccess the application at http://localhost:8080.
- Frontend: Next.js 15+ with Tailwind CSS and Framer Motion for a premium, responsive UI.
- Backend: Rust (Axum) featuring:
- AST Parser: Regex-based hierarchical parsing of legal documents.
- Aligner: Multi-stage alignment logic to handle document evolution.
- Tokenizer: Efficient Chinese segmentation using
jieba-rs.
- Proxy: Caddy for efficient static file serving and API reverse proxying.