There’s no direct equivalent function. Instead, PostgreSQL lets you get the generated id back from the INSERT itself, with a RETURNING clause — one round trip, no separate lookup call needed.

The MySQL pattern

INSERT INTO orders (customer_id) VALUES (42);
SELECT LAST_INSERT_ID();

Two statements, and — worth knowing even before you migrate — LAST_INSERT_ID() is connection-scoped, so it’s already fragile around connection pooling if you’re not careful about which connection issued the insert.

The PostgreSQL way: RETURNING

INSERT INTO orders (customer_id) VALUES (42)
RETURNING id;

One statement, one round trip, and the returned id is unambiguously tied to the row you just inserted — no connection-scoped state to reason about. RETURNING isn’t limited to id, either: return any column, or several, or RETURNING * for the whole row.

If you’re not doing the INSERT yourself

Working through a driver or ORM instead of raw SQL, look for whichever call actually executes an INSERT ... RETURNING under the hood — most PostgreSQL drivers and ORMs use this pattern already for “give me back the row I just created,” you’re just not writing the SQL by hand. If you’re stuck without RETURNING for some reason, currval('sequence_name') is the direct sequence-based equivalent — see the AUTO_INCREMENT / sequences FAQ for how PostgreSQL sequences map to MySQL’s AUTO_INCREMENT in the first place.