PostgreSQL

Instantly wipe every row from a PostgreSQL table in a single atomic operation, leaving the schema intact and ready for fresh data.

Insert a single new record into any PostgreSQL table, mapping incoming data fields to the target columns.

Reset one or more PostgreSQL auto-increment sequences back to their starting value, keeping ID counters consistent after a truncate or bulk delete.

Modify one or more fields in an existing PostgreSQL row identified by a column match, without touching any other records.

Remove specific rows from a PostgreSQL table that match your filter criteria, without touching the rest of the table.

Look up a single PostgreSQL row by matching one or more column values and return the full record for downstream use.

Retrieve multiple rows from a PostgreSQL table using column filters, returning a result set the agent can loop over or forward downstream.

Run a hand-written SELECT statement against PostgreSQL and return all matching rows, giving the agent full SQL expressiveness for complex multi-table lookups.

Compose AND / OR filter logic across multiple columns so the agent can target precisely the rows it needs without writing raw SQL.

Execute a custom SQL SELECT that returns a single matching row, ideal when the lookup requires a subquery, CTE, or computed column unavailable in standard filters.

Permanently remove a PostgreSQL table, index, or view from the schema — use with care, as this cannot be undone without a restore.

Send one SQL statement per database call — the default mode, guaranteeing each statement's result is available before the next action runs.

Execute multiple SQL statements in a single network call, each running independently so a failure in one does not roll back the others.

Wrap multiple SQL statements in a single BEGIN / COMMIT block so all succeed together or all roll back — essential for operations that must stay atomic.

Return NUMERIC and BIGINT column values as native JavaScript numbers, suitable when downstream steps expect arithmetic-ready data and the values fit safely within a 64-bit float.

Return NUMERIC and BIGINT column values as strings, preventing precision loss for very large integers or high-precision decimals that exceed JavaScript's safe integer range.

Explicitly map each incoming data field to its target PostgreSQL column by name, giving full control over which values land where.

Let the agent match incoming data fields to PostgreSQL columns automatically by name, ideal when the source payload and table schema share the same field names.

Fires via PostgreSQL LISTEN/NOTIFY the moment a new row is inserted into the watched table, letting the agent react in real time without polling.

Fires when a new column is added to any monitored table, allowing the agent to react to schema evolution as it happens.

Fires via LISTEN/NOTIFY whenever a row in the watched table is updated, letting the agent react to state changes without polling the database on a schedule.

Polls the watched table on a schedule and fires once for each row that did not exist on the previous check, covering tables that cannot use LISTEN/NOTIFY.

Fires via LISTEN/NOTIFY when a row is deleted from the watched table, enabling the agent to cascade actions or audit deletions in real time.

Polls using a hand-written SELECT and fires for each row returned that is new since the last check, allowing complex multi-table or windowed conditions to define what counts as a new event.

Fires when a new table is created in the monitored PostgreSQL schema, letting the agent respond to schema expansion events automatically.

Fires for both newly inserted and recently modified rows in the watched table, covering any change to a record whether it arrived fresh or was edited in place.