Vector Database Toolkit

Unified VectorStore interface with 5 backends (In-Memory, Pinecone, Weaviate, ChromaDB, pgvector), benchmarking harness, and migration tool.

Select a backend, configure the collection, insert sample vectors, and search.

Backend Dimension Index Type Sample Vectors Search Query (text)

Search Results

Configure and click to search.

Latency & Recall (top-10, 100K vectors)

In-Memory: 2.3ms / recall@10: 1.000
HNSW (ef=128): 0.4ms / recall@10: 0.985
IVFFlat (lists=100): 0.8ms / recall@10: 0.942

Full version includes common interface, 5 adapters, in-memory reference, benchmarking harness, migration tool, and indexing guide.

🤖 AI/LLM Toolkit — Vector Database Toolkit Demo