Nexura Qdrant

A Qdrant vector store with workspace-scoped collections and immutable index generations.

Install

pip install nexura-qdrant

Alpha (0.4): pip install nexura-qdrant. See package installation.

Minimal example

Construction is lazy; this example refuses zero dimensions before any request. A running Qdrant is required for create_collection, upsert and search. Workspace keys scope records and generations; active index generations are immutable. Built-in runtime vector stores work without this add-on.

import asyncio

from nexura_runtime.knowledge.vector.errors import VectorDimensionMismatchError
from nexura_runtime.knowledge.vector.store import VectorStore

from nexura_qdrant import QdrantVectorStoreAdapter


async def main() -> None:
    store = QdrantVectorStoreAdapter(url="http://localhost:6333")
    assert isinstance(store, VectorStore) and store.is_available
    try:
        await store.create_collection("ws-1", "faq", dimensions=0)
    except VectorDimensionMismatchError as exc:
        print("refused before any request:", exc)
    await store.close()


asyncio.run(main())

Status and limits

Status: alpha (0.4), Python 3.12 (the same as nexura-runtime). Tested against a local disposable Qdrant service; never against Qdrant Cloud.

Documentation and licence

Main documentation · Full guide. MIT, Copyright (c) 2026 Liceron. See LICENSE.