Point-in-time recovery, often shortened to PITR, lets you restore a database to its state at a specific moment, not just to the time of a fixed backup. It works by combining a base backup with a continuous record of the changes made afterwards, often called write-ahead or transaction logs, which can be replayed up to the chosen time.
It is most useful for mistakes and corruption, such as an accidental delete, a bad migration or a faulty application release. You pick a time just before the problem and restore to that point, which can lose far less data than going back to last night's backup.
PITR has limits. It only reaches back as far as the retention window, which differs between services and plans, and it does not help if the change logs were not being kept. Restores often create a new instance rather than overwriting the existing one, so you need to plan how to switch over and check the result. It is also not a substitute for separate backups stored elsewhere.
Check what your provider offers, how long the history is kept, how long a restore takes and what it costs. Test a restore before you need one, so the process is familiar when something goes wrong.