ai-rag

What problem does it solve?

Automates building production-grade retrieval-augmented generation systems, reducing manual integration work, tuning, and evaluation effort for real-world knowledge apps.

Core Features & Use Cases

• Chunking & grounding: Page-level and semantic chunking for accurate grounding.
• Hybrid retrieval: BM25 lexical + vector search with reranking.
• RAG evaluation: Recall@K, nDCG, groundedness metrics for quality validation.
• Production patterns: Context injection, prompt caching, and grounded generation.

Quick Start

Use ai-rag to initialize a retrieval-augmented generation pipeline for a knowledge base and test grounding with a sample query.