Shiny.DocumentDb v6 — Vectors, Filters & a Real Connection Pool | Community Feed

Shiny.DocumentDb v6 closes five major gaps that have been on the wish list since v3. The same zero-schema document model and AOT story remain, but v6 adds capabilities that make it viable for real ASP.NET Core services alongside its existing mobile strengths.

What you’ll learn

Vector/ANN search — one API that translates to pgvector, SQL Server 2025 DiskANN, Cosmos DB, MongoDB Atlas, DuckDB vss, and sqlite-vec
Global query filters — AddQueryFilter<T>() that enforces predicates on every read path, matching EF Core’s HasQueryFilter semantics
Composite JSON indexes — multi-expression index creation across all relational providers
Connection pooling — server SQL providers (PostgreSQL, MySQL, SQL Server) now use ADO.NET driver pools instead of a single in-process semaphore
Per-query change monitoring — .NotifyOnChange() on fluent queries filtered by their own Where predicates
Auto-embedding on insert with Shiny.DocumentDb.Extensions.AI