sqlite-vec

Brinicle is a lightweight C++ library for approximate nearest neighbor (ANN) vector search on embeddings, optimized for low-RAM environments rather than full vector databases. It features efficient graph-based indexing (HNSW-like), supports quantization for further memory reduction, and excels in languages like C++. Ideal for rapid prototyping of ML prototypes and embedded applications; lighter and more memory-efficient than Milvus, with better low-resource performance vs hnswlib.

LEANN is a lightweight RAG-focused library for vector search on embeddings, achieving 97% storage savings via advanced compression and quantization techniques on personal devices. Implemented in Rust/Python, it supports efficient ANN indexing without full DB overhead. Ideal for embedded apps and private prototyping; far lighter than Milvus, more efficient on-device vs hnswlib.

Haystack is a Python library for building vector search and embedding-based retrieval pipelines, integrating ANN indexes without requiring full databases. Key features include support for HNSW, FAISS indexes, quantization options, and multi-language embeddings. Perfect for prototyping RAG systems and embedded AI apps; more flexible than hnswlib, lighter than Milvus for development workflows.

LangChain4j is a Java library providing vector search and embedding capabilities for LLM applications via integrations with ANN indexes like HNSW and FAISS, without needing full vector databases. Features include support for quantization, tool calling, and seamless embedding in JVM environments like Spring Boot and Quarkus. Suited for prototyping RAG agents and embedded apps; lighter and more JVM-native than Milvus, easier integration vs hnswlib.

LlamaIndex is a Python data framework library for vector search and embedding retrieval, integrating various ANN indexes like HNSW and FAISS without full database dependencies. Supports quantization, multi-modal embeddings, and advanced query engines in Python/Rust backends. Great for prototyping LLM apps and embedded RAG; more developer-friendly and lighter than Milvus, composable vs hnswlib.

Multilingual embedding model supporting over 100 languages and ranking #1 on MTEB multilingual leaderboard. Offers flexible model sizes from 0.6B to 8B parameters with user-defined instructions.