OpenCrawling integrates Qdrant for secure enterprise RAG
TL;DR. OpenCrawling developed a Qdrant Output Connector for enterprise RAG, focusing on secure vector search with document-level access controls. - The connector uses Qdrant's Payload Index Engine to filter by security SIDs before distance calculations. - It supports decoupled ingestion pipelines and scalable embedding services like Ollama or OpenAI for vector creation. - Automated collection provisioning and ACL indexing ensure that user access rights are enforced on vector points.
- Qdrant Output Connector enables secure enterprise RAG with document-level ACLs.
- Utilizes Qdrant's Payload Index Engine for pre-filtering security SIDs in vector search.
- Integrates into OpenCrawling's microservice architecture, supporting various embedding services.
- Automates Qdrant collection setup and payload ACL indexing for read permissions.
Sources
- The Qdrant Output Connector — opencrawling.org