Connect to PostgreSQL database.

Execute a SELECT query

Validate SQL. Returns the string 'true' if the SQL is valid, or an error message string if invalid.

Trigram search. Resolve mentioned entities to database matches. Returned rows are candidates, not confirmed matches.

Execute an INSERT, UPDATE, or DELETE query

List all schemas in the database

List tables in the database

Get table structure for single or multiple tables

List tables that have pgvector columns (vector or halfvec), with schema, column, and data type.

Nearest-neighbor search. Embeds each query string and ranks with pgvector. query is an array: one element runs one search; multiple elements run the same search (same table, where, columns, limit) once per element and return grouped results. Pass text_column to fuse each ranking with a lexical list using reciprocal rank fusion (k=60). lexical=fts (default) uses English full text: stems, drops stop words, keeps Arabic, and drops query terms that appear in more than 40% of rows. lexical=trgm uses pg_trgm word_similarity above pg_trgm.word_similarity_threshold. Omit text_column for dense search only. A where clause sets hnsw.iterative_scan=strict_order for that query. hnsw.ef_search is raised when the candidate limit is above 40. Default payload omits embedding and *hash. include_scores defaults to true and adds distance, plus rrf, dense_rank, and lexical_rank when hybrid. Set include_scores false to omit those columns. response_format=markdown returns one markdown table per query element.

Insert or update a vector/halfvec column by embedding text. insert: embed text (optional fields, optional persist via text_column). update: write only the vector. Source is text (single row: id, or where that matches one row) or text_column (embed each matched row, where/id, optional limit). Use execute for delete and non-vector columns.