Qdrant Vector Memory

Use Qdrant for vector search, semantic caching, and long-term agent memory. Builds collections with metadata filtering, hybrid search, and semantic caches that cut LLM calls and shrink context windows.

This skill builds vector-search and memory layers on Qdrant. It designs collections and payload schemas, combines vector similarity with keyword and metadata filtering for hybrid search, implements semantic caching to skip repeat LLM calls, and stores long-term agent memory so retrieval returns only the relevant context chunks.

When to use

Use when adding a vector database, semantic cache, or long-term memory to an LLM app — with metadata filtering and hybrid search over embeddings.

Examples

Build a vector store

Collection with filtering

Create a Qdrant collection for product embeddings with payload filters on category and price, and a hybrid search query

Add semantic caching

Cut redundant LLM calls

Implement a Qdrant-backed semantic cache that returns a stored answer when a new query is sufficiently similar to a past one