Glossary

What is reranking?

Reranking reorders an initial set of retrieval results with a more accurate — and costlier — model, putting the most relevant items on top.

Retrieval is usually two stages: a fast retriever pulls a broad candidate set, then a reranker scores each candidate against the query with a more precise model and reorders them. This sharply improves the quality of the context fed into a RAG system, which directly improves its answers.

Open-source cross-encoder rerankers drop into existing vector-search pipelines as that second stage.

Best vector databases →

Trending reranking projects

Tencent/WeKnora
Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
★ 15.7K+5 · 24hmomentum 14Go
VectifyAI/PageIndex
📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG
★ 32.3K-6 · 24hmomentum 8Python
memvid/memvid
Memory layer for AI Agents. Replace complex RAG pipelines with a serverless, single-file memory layer. Give your agents instant retrieval and long-term memory.
★ 15.6K-1 · 24hmomentum 8Rust
yichuan-w/LEANN
[MLsys2026]: RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.
★ 11.8K0 · 24hmomentum 7Python

▌ reranking — FAQ

What is reranking?

Reranking reorders an initial set of retrieval results with a more accurate — and costlier — model, putting the most relevant items on top. Retrieval is usually two stages: a fast retriever pulls a broad candidate set, then a reranker scores each candidate against the query with a more precise model and reorders them. This sharply improves the quality of the context fed into a RAG system, which directly improves its answers.