Show HN: Pg_CRDT – CRDTs in Postgres Using Automerge (opens in new tab)

And could this be used as backend support for https://github.com/drewmccormack/Forked?

We just released a new version which solves this problem: https://github.com/supabase/pg_crdt

The new approach uses UNLOGGED tables for documents and logged tables for changes. This enables us to merge changes efficiently in-memory, then persists just the changes directly to the WAL (previously it was sending the entire CRDT).

The new version also takes advantage an advanced in-memory object persistence feature in Postgres called "expanded datum"[1] and some optimizations to this feature that are coming in Postgres 18[2]. This allows for more complex operations without serializing and deserializing the documents between every operation, which gets progressively slower as documents grow.

What is a CRDT? A CRDT (Conflict-free Replicated Data Type) is a data structure that syncs across multiple devices or users. They are used in multi-player applications like Figma and Notion. For example: A shared text editor where multiple people type at once, and everyone sees the same final result without manual conflict resolution.

The end goal here is that you will be able to store CRDTs as a data type in your database, and treat Postgres simply as a "peer". A good example would be storing text for a blog post:

    create table posts (
      id serial primary key,
      title text not null
      content autodoc not null default '{}'
    );

Docs: https://supabase.github.io/pg_crdt/automerge/

----

[0] initial release/discussion: https://news.ycombinator.com/item?id=33931971

[1] expanded datum: https://www.postgresql.org/docs/current/storage-toast.html#S...

[2] PG18 optimizations: https://www.postgresql.org/message-id/3363452.1737483125%40s...