ManyVector

What ManyVector is

ManyVector is a vector database built on object storage, with vector, full-text, and hybrid search for uses such as RAG pipelines, semantic search, recommendations, and AI agent memory. It uses HNSW approximate nearest-neighbour search and advertises sub-10ms p50 latency with 90%+ recall@10. BM25 supports full-text search; hybrid search combines vector and text scores through RRF and metadata filters. Copy-on-write namespace branching allows snapshots, evaluation branches, and rollback without copying vector data. Supported object stores include AWS S3, Google Cloud Storage, Cloudflare R2, Azure Blob Storage, MinIO, Tigris, Wasabi, and S3-compatible stores. ManyVector says it works with OpenAI, Cohere, and other embedding models or providers; the Free Tier includes Python and TypeScript SDKs. The company says vectors and indexes remain in the customer's object store rather than on its own disks. Managed SaaS traffic uses TLS 1.2 or higher, and data at rest is encrypted with AES-256. The Free Tier includes 100M search queries per month and requires a customer-provided bucket. Growth includes 100M free queries per month, with billing per query above that allowance. Enterprise options include self-hosted deployment and dedicated support.