Dev Tools · 1h ago
Ukraine's 33.7M court decisions vectorized for AI legal search
OverTheLex vectorized 33.7M Ukrainian court decisions using Voyage AI's embedding model, storing 44M+ vectors in Qdrant. The pipeline runs on a single EC2 instance and processes civil cases at 42% completion. The system enables semantic search for lawyers, replacing keyword-based full-text search.
Meridian48 take
The scale is impressive but the real test is whether lawyers actually adopt this over traditional search; the 63M vector target suggests serious infrastructure costs ahead.
Read the full reporting
How We Vectorize 33.7M Ukrainian Court Decisions via Voyage AI →
DEV Community
vector-databaselegal-tech