Grep is All You Need: Zero-Preprocessing Knowledge Retrieval for LLM Agents

Retrieval-Augmented Generation (RAG) has become the dominant paradigm for grounding Large Language Model (LLM) agents in domain-specific knowledge. The standard approach requires selecting an embedding model, designing a chunking strategy, deploying a vector database, maintaining indexes, and performing approximate nearest neighbor (ANN) search at query time. We argue that for domain-specific knowledge grounding — where the vocabulary is predictable and the corpus is bounded — this entire stack is unnecessary. We present *Knowledge Search*, a two-layer retrieval system composed of (1) `grep` with contextual line windows over raw source texts and (2) `grep` over LLM-compiled per-source concept and FAQ files generated nightly by a free, local, autonomous compilation pipeline. Deployed in production across **76 specialized LLM agents** serving three knowledge domains (Traditional Chinese Medicine, Christian spiritual classics, U.S. civics) — grounded in approximately **500 primary source texts and ~180 MB of corpus** spanning two languages and four-and-a-half millennia of human thought, served by a single Mac mini — our approach achieves 100% retrieval accuracy with sub-10ms latency, zero per-query preprocessing, zero additional memory footprint, and zero infrastructure dependencies. We also document a reproducible failure-and-recovery cycle (0/5 fabricated quotes → 4/4 grep-verified quotes after a 25-minute fix that touched only text files on disk) which demonstrates the architecture's safety properties are recoverable through prompt hygiene alone — no retraining, no infrastructure change. The key insight is simple: retrieval does not need intelligence. The LLM is the intelligence. Bilingual paper (English + 中文). Production system: https://faith.localkin.ai · https://heal.localkin.ai. Code: https://github.com/LocalKinAI/grep-is-all-you-need.