Hora

ClickHouse is a columnar OLAP database with vector indexes (ANN via AMM, brute-force), supporting SQL queries over vectors + structured data at petabyte scale. Excels in aggregations with vectors. For analytics workloads with embeddings; faster ingestion than Postgres pgvector for big data.

nanopq is a lightweight product quantization library for efficient vector compression and similarity search, which is an important feature for vector databases that need to store and query large-scale vector data efficiently.

Local LLM inference engine supporting GGUF models with hardware acceleration on Metal, CUDA, ANE, WebGPU. Features Flash Attention, MicroLoRA, RoPE, quantization (Q4-Q8, π-Quantization), MoE routing, and streaming tokens for browser and edge deployment.

Open-source project demonstrating GPU-accelerated approximate nearest neighbor search using Inverted File (IVF) indexing on embeddings from a large Wikipedia dataset. It employs K-means clustering into 128 clusters and supports configurable CUDA kernels for coarse and fine search stages. Applicable for efficient vector querying in AI applications.

Lucene KNN vector plugin for Elasticsearch search engine, enabling hybrid lexical+vector search, BM25 fusion, HNSW/IVF indexes for ANN. Used for enterprise search, RAG, multimodal apps. Integrated vs standalone like Weaviate: superior hybrid text handling but higher resource footprint.

Embeddable SQL OLAP engine with VSS extension for low-latency HNSW vector search on local files, ideal for edge AI prototyping and analytics. SQL-first approach for on-device vector ops vs cloud vector DBs like Qdrant.