thustorage/PipeANN
A low-latency, billion-scale, and updatable graph-based vector store on SSD.
Implements a pipelined SSD-based vector search using an in-memory entry point index and graph-based layout with PQ compression, achieving 20K QPS for 1B vectors while supporting arbitrary filtered search via user-defined selectors. Built on `io_uring` for async I/O and compatible with existing DiskANN indexes, with both Python (faiss-like API) and C++ interfaces for search, insert, and delete operations with minimal performance degradation during updates.
102 stars.
Stars
102
Forks
37
Language
Jupyter Notebook
License
MIT
Category
Last pushed
Feb 04, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/vector-db/thustorage/PipeANN"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related tools
MariaDB/server
MariaDB server is a community developed fork of MySQL server. Started by core members of the...
AlayaDB-AI/AlayaLite
AlayaLite – A Fast, Flexible Vector Database for Everyone.
infiniflow/infinity
The AI-native database built for LLM applications, providing incredibly fast hybrid search of...
gusye1234/nano-vectordb
A simple, easy-to-hack Vector Database
nnethercott/hannoy
Production-ready KV-backed HNSW implementation in Rust using LMDB