NemoClaw Knowledge Wiki

agentic-search

2 min read

🗂️ AI & Agents · View mindmap

Agentic Search

Agentic search is an information retrieval approach where autonomous agents actively explore file systems or knowledge bases to locate relevant information, rather than relying on static, pre-indexed retrieval methods. In this model, an agent navigates through documents and directory structures iteratively, evaluating relevance based on discovered content and deciding which paths to explore next. This transforms information discovery from a lookup operation into an exploration process, allowing agents to adapt their search strategy based on what they encounter.

Distinction from RAG

Traditional retrieval-augmented generation (RAG) relies on vector similarity search and pre-computed embeddings to retrieve relevant documents before generating responses. Agentic search differs by treating retrieval as a dynamic process where agents make sequential decisions about which files or directories to examine. Rather than computing similarity scores against an entire indexed corpus, agentic approaches allow agents to navigate hierarchies, follow logical connections, and progressively narrow or expand their search scope based on intermediate findings.

Applications and Considerations

The fs-explorer project demonstrates agentic search applied to file system exploration, where agents systematically navigate directory structures to fulfill information requests. This approach can be useful when dealing with unstructured or evolving knowledge bases where pre-indexing is impractical or where the search space benefits from contextual navigation. However, agentic search may require more computational steps than direct retrieval, and its effectiveness depends on how well agents can interpret directory structures and document content to make intelligent exploration decisions.

Source Notes

Graph View

Backlinks