About BabelCite
A professional legal research platform built to make Hong Kong case law fast, searchable, and genuinely useful for lawyers and researchers.
Our Mission
Research that keeps up with how lawyers actually think
Keyword search was designed for librarians, not litigators. Lawyers reason about legal concepts, factual parallels, and analogous reasoning — not exact phrase matches.
BabelCite bridges that gap: describe your issue in plain English, and the AI surfaces the cases that genuinely matter — ranked by why they match, not just that they do.
Concept-first search
AI understands the legal principle you're researching, not just the words you typed.
Transparent relevance
Every result is labelled — similar issue, similar facts, or similar outcome — so you know exactly why it appeared.
Structured summaries
AI-generated headnotes extract key facts, legal issues, and outcome for quick triage.
Full original text
Read the complete judgment with clean formatting and paragraph anchors. Nothing omitted.
Coverage
All major Hong Kong courts
Hybrid search
Semantic vector search (Qdrant) combined with BM25 keyword search (SQLite FTS5) for precision across both concept and citation.
Citation graph
185,000+ resolved citation edges across the corpus. Importance scoring identifies leading authorities across all courts.
AI enrichment
Top cases are enriched with structured headnotes, legal issues, outcome, and judge — extracted by large language models.
How It's Built
Purpose-built for legal research
The entire Hong Kong judiciary database — over 180,000 decisions — is indexed in a hybrid search architecture combining dense vector embeddings and sparse keyword matching. Results are fused and re-ranked by AI.
A citation graph maps cross-references between cases, enabling importance scoring that surfaces leading authorities — not just recent ones.