Skip to Content
Agentic WorkspaceToolsBuilt-in toolsRAG & knowledge

RAG & knowledge

Search and manage your project’s content stores — vector indexes of documents, crawled pages, and other sources that agents search semantically.

Semantic/hybrid search over content stores; returns relevant chunks and citations. Category: Data Pull · Setup: needs a populated content store.

  • Inputs: query (required), store_id (optional — omit to search all bound stores), max_results (default 5), similarity_threshold (default 0.3), source_type, include_metadata.
rag_search tool — Overview tab
rag_search tool — Test tab returning chunks

rag_list_stores

List available content stores (IDs, status, counts). Category: Data Pull · Setup: none (empty list if none exist).

  • Inputs: status_filter, include_stats.
rag_list_stores tool — Overview tab
rag_list_stores tool — Test tab

rag_get_store_status

Status and stats of a store, including indexing progress and errors. Category: Data Pull · Setup: needs a store.

  • Inputs: store_id (required).
rag_get_store_status tool — Overview tab
rag_get_store_status tool — Test tab

rag_reindex

Reprocess a store’s documents and regenerate embeddings. Category: Action · HIL approval · Setup: needs a store.

  • Inputs: store_id (required), full_reindex, document_ids.
rag_reindex tool — Overview tab
rag_reindex tool — Test tab
ℹ️

rag_search returns “No content stores found for this project. Create a store first.” if none exist. Use rag_list_stores to discover store IDs. rag_reindex pauses for approval because reindexing is expensive.

DatabaseDocuments & storage