qdrant-client

Why we included this project

If your project leans on Qdrant for semantic search or retrieval-augmented generation, this Python client is where most of your day-to-day work will happen. It ships typed models and helper methods for every Qdrant API, so you can create collections, upsert points in bulk, and run filtered similarity queries without hand-writing raw requests. The sync and async variants, plus REST and gRPC support, let you match the client to your application's concurrency style. A genuinely useful local mode runs the same code against an in-memory or on-disk store, so you can prototype in a notebook or test in CI before pointing identical calls at a real server. Recent additions also bundle local CPU embedding via FastEmbed, which shortens the distance between raw documents and a working search index.