turbopuffer
turbopuffer 把向量检索与全文搜索建在对象存储上：向量索引与倒排索引都放在 S3 这类廉价存储里，冷数据不占内存，查询按命名空间读取，天然适合每个客户开一个独立命名空间的多租户产品。接口同时支持向量近邻、BM25 关键词检索与两者混合排序，官方文档给了 Python、TypeScript、Go 的客户端示例。 🔗 文档: https://turbopuffer.com/docs
turbopuffer builds vector search and full-text retrieval on top of object storage: both the vector index and the inverted index live in cheap storage such as S3, cold data stays out of memory, and queries read per namespace, which suits multi-tenant products that give every customer an isolated namespace. The API covers vector nearest-neighbour, BM25 keyword search and hybrid ranking, with Python, TypeScript and Go clients in the docs. 🔗 Docs: https://turbopuffer.com/docs
📋 Product Info
🎯 Why This Tool
turbopuffer belongs to the Developer Tools category on hedirbase. Tagged with: 向量数据库,全文检索,对象存储,无服务器.
turbopuffer builds vector search and full-text retrieval on top of object storage: both the vector index and the inverted index live in cheap storage such as S3, cold data stays out of memory, and que