Index Legal Docs for Hybrid Search: Qdrant, OpenAI & BM25

This n8n workflow indexes a legal Q&A corpus from Hugging Face (isaacus/LegalQAEval) into Qdrant, enabling hybrid search that combines semantic similarity via dense vectors and keyword matching via sparse vectors (BM25). It supports two embedding options: Qdrant Cloud Inference for direct vectorization or external providers like OpenAI's text-embedding-3-small or mxbai-embed-large-v1. After execution, your Qdrant collection is ready for advanced retrieval in the companion workflow.

Key benefits include seamless setup for Legal AI applications, reduced manual data processing, and scalable hybrid search outperforming pure semantic or keyword methods. It saves hours of custom scripting by automating dataset loading, embedding generation, and indexing with configurable parameters for cluster credentials and model selection.

Ideal use cases: Building RAG systems for legal research, contract analysis, compliance tools, or any domain needing precise document retrieval. Prerequisites include a Qdrant Cloud cluster (paid US region for inference). Part of a series on Hybrid Search with Qdrant & n8n for Legal AI.