ragleap-vectorstores

Pluggable vector backends beyond ragleap-rag's built-in 6 (PgVector, FAISS, Pinecone, Weaviate, Qdrant, Milvus).

This package is under active development - more backends will be added over time. See the roadmap for status.

Available backends

Design

Every backend here implements ragleap-rag's VectorBackend interface, so it can be passed directly to RagLeap(vector_backend=...). Each backend's real client SDK is an optional extra - installing ragleap-vectorstores alone pulls in no heavy dependencies beyond ragleap-rag itself.

Install

pip install ragleap-vectorstores[chroma]
# or, with uv
uv add ragleap-vectorstores[chroma]

pip install ragleap-vectorstores[lancedb]
# or, with uv
uv add ragleap-vectorstores[lancedb]

pip install ragleap-vectorstores[opensearch]
# or, with uv
uv add ragleap-vectorstores[opensearch]

Usage

from ragleap_vectorstores import ChromaBackend

backend = ChromaBackend(persist_directory="./chroma_data")

from ragleap_vectorstores import LanceDBBackend

backend = LanceDBBackend(uri="./lancedb_data")

from ragleap_vectorstores import RedisBackend

# Requires a real Redis Stack instance (or plain Redis + the RediSearch
# module) - plain Redis alone has no vector search.
backend = RedisBackend(redis_url="redis://localhost:6379/0")

from ragleap_vectorstores import UpstashBackend

# Requires a real Upstash Vector index, created via the Upstash console
# (console.upstash.com) as a pure DENSE index - no embedded/local mode.
backend = UpstashBackend(url="https://...upstash.io", token="...")

from ragleap_vectorstores import OpenSearchBackend

# Requires a real running OpenSearch instance (e.g. http://localhost:9200) -
# no embedded/local mode. Pass http_auth=("user", "password") and
# use_ssl=True for a secured cluster.
backend = OpenSearchBackend(opensearch_url="http://localhost:9200")