PostgreSQL: Non-continious replication of new data with COPY

I have a database tables with millions of measurements. New data is coming in every day. For analysis, I want to replicate the data to my laptop (local postgres db). No need for automatic replication. It's OK to start a script for this. I feel like the standard replication solutions are inadequate and overengineered for my use because - I need asynchron replication, - some solutions are inefficient because they replicate row-wise (yet data on the server is bulk-inserted), - and they need too much configuration. I would like to use COPY TO STDOUT | COPY FROM STDIN but here I have to select only the new data. How can I do that? The table has this form: Column | Type ------------+-------------------------- devicename | character varying(30) id | integer timestamp | timestamp with time zone value | numeric variable | text PK would be (devicename, id). Note that id alone is not unique because the data is coming from multiple devices. How can I select only new data for COPY? Any other approaches for this replication requirements?