NeMo Retriever MCP
Use the NeMo Retriever MCP tools exposed under retriever instead
of the retriever CLI.
The server exposes MCP tools, not MCP resources. Do not call
list_mcp_resources to discover documents or decide whether Retriever is
available; an empty MCP resource list is expected.
Workflow
- Use
queryto search documents already available to Retriever. - Never ingest documents that are already searchable, including documents in a prebuilt index.
- Use
ingest_documentsonly when required documents are not yet searchable and server-readable paths are explicitly available. Do not assume that paths visible to the agent are also readable by the Retriever server. - After ingestion completes, use
queryto retrieve evidence from the added documents. - Start with
top_k=5andformat="hits". Request more only when the first successful response does not contain enough evidence to answer the task. - Read evidence hits from the tool response and ground your answer in them.
- Preserve available document identifiers such as
document_id,source_id, orpdf_basename, along with page numbers, when the task requests citations or structured output. - If
queryfails, inspect the error and retry with corrected parameters or a shorter equivalent query. If it still fails, report the error before using another retrieval method.