LanceDB

LanceDB began as an embedded, open source vector database built on the Lance file format, designed for vector, text, and hybrid search without server administration. The library (Python, JavaScript, Rust) remains free and self-hostable without limits, locally or on your own object storage (S3, GCS, Azure Blob). The vendor has, however, redirected its commercial messaging toward a broader multimodal lakehouse platform intended for large-scale model training, and has removed all public pricing from its website: the pricing page now offers only a contact form for both LanceDB Cloud and the Enterprise offering. No amounts, pricing tiers, or indicative quotes are communicated without engaging with the sales team.

LanceDB is available only in English, on the website as in the documentation. The vendor does not publish GDPR commitments or publicly documented European hosting; this must be negotiated on a case-by-case basis during sales outreach. The open source library has a complete API and SDKs, with documented integrations to LangChain, LlamaIndex, OpenAI, Cohere, Hugging Face, and DuckDB. The most critical limitation for a small organization is the complete absence of published pricing beyond the self-hosted open source version: there is no way to know whether the managed service will cost tens or thousands of dollars per month without going through a commercial sales cycle.