Convert and index one document from a configured document root.

path may be relative to RAG_DOCUMENT_ROOT or an absolute path. Supports: PDF, DOCX, XLSX, CSV, PPTX, HTML, Markdown, TXT, XML. Returns validation results and diff info (skipped if unchanged). Set force=True to force re-indexing even if content hash is unchanged. Set async_index=True to run in background and return job_id.

Index supported files in a directory with diff reporting.

Returns which files are new, changed, unchanged, or failed. Set force=True to force re-indexing all files. Set async_index=True for large directories to prevent timeout.

Search indexed documents with hybrid dense+sparse retrieval.

Optionally filter by metadata (e.g. filter={"sheet_name": "Branch_Alpha"}). Results include structured citations with sheet name and row ranges. Cross-encoder reranking is applied when configured.

Execute a SQL query on extracted tables stored in DuckDB.

Use list_tables() first to discover available table names and column schemas. Supports COUNT, GROUP BY, JOIN, WHERE, etc.

Simplified aggregation on a DuckDB table.

Easier than SQL for common count/sum/group-by queries. Example: aggregate(table="t_abc_0", group_by=["xã"], agg="count")

List all extracted tables in the DuckDB relational store.

Returns table names, source documents, column schemas, and row counts.

List indexed files, hashes, chunk counts, timestamps and source paths.

Return all indexed chunks and provenance for one document id.

Return Qdrant/DuckDB health, counts, roots and dependency versions.

Return a read-only health snapshot for this MCP and its backends.

Return safe read-only runtime configuration without secrets.

Check the status of an async indexing job.